How Well Can Current Climate Models Simulate the Connection of the Early Spring Aleutian Low to the Following Winter ENSO?

Shangfeng Chen aCenter for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China

Search for other papers by Shangfeng Chen in
Current site
Google Scholar
PubMed
Close
,
Wen Chen aCenter for Monsoon System Research, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China

Search for other papers by Wen Chen in
Current site
Google Scholar
PubMed
Close
,
Bin Yu bClimate Research Division, Environment and Climate Change Canada, Toronto, Ontario, Canada

Search for other papers by Bin Yu in
Current site
Google Scholar
PubMed
Close
, and
Renguang Wu cSchool of Earth Sciences, Zhejiang University, Hangzhou, China

Search for other papers by Renguang Wu in
Current site
Google Scholar
PubMed
Close
Free access

Abstract

A recent study revealed an impact of the intensity of early spring Aleutian low (AL) on the succeeding winter ENSO. This study examines the ability of 41 climate models that participated in CMIP6 in simulating the early spring AL–winter ENSO connection. It is shown that there exists a large diversity among the models in simulating this AL–ENSO linkage. A number of models capture well the observed AL–ENSO connection and the associated physical processes. However, the AL–ENSO relation in several models is opposite to the observed. Diversity of the AL–ENSO connection is related to the spread in the spatial structure of AL-related atmospheric anomalies over the North Pacific. In the models that capture the observed AL–ENSO connection, weakened AL induces an anomalous anticyclone over the northern middle and high latitudes and an anomalous cyclone over the subtropical North Pacific. The resultant westerly wind anomalies over the tropical western-central Pacific (TWCP) induce an El Niño sea surface temperature (SST) anomaly pattern in the following winter. By contrast, in the models with the AL–ENSO relation opposite to the observations, the AL-associated anomalous anticyclone over the North Pacific extends too southward. As such, the subtropical North Pacific is dominated by northeasterly wind anomalies and SST cooling. The subtropical North Pacific SST cooling induces easterly wind anomalies over the TWCP via wind–evaporation–SST feedback, and leads to a La Niña anomaly pattern in the following winter. The spread in the spatial structure of the AL-associated atmospheric anomalies over the North Pacific is partly due to the diversity in the amplitude of the climatological mean flow.

Significance Statement

A recent study suggested that variation of the AL intensity in early spring could exert a significant impact on the following winter ENSO. It indicated that inclusion of the early spring AL signal could improve the prediction of ENSO and to some extent help reduce the spring predictability barrier of ENSO. To employ the AL as a predictor in the ENSO prediction and forecast, the current climate model should have the ability to simulate realistically the early spring AL variation as well as the physical process linking the early spring AL with the subsequent winter ENSO. Hence, this study examines the performance of the current coupled climate models that participated in the phase 6 of the Coupled Model Intercomparison Project (CMIP6) in simulating the linkage between the early spring AL and the following winter ENSO. We show that there exists a large diversity among the CMIP6 models in simulating the early spring AL–winter ENSO connection. A number of models capture well the observed AL–ENSO connection and the associated physical processes. However, the AL–ENSO relation in several models is opposite to the observed. The factors leading to the spread are further examined. Results of this study would have implications in improving our understanding of the impact of extratropical atmospheric forcing on the ENSO and improving the seasonal forecasting of the ENSO.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Shangfeng Chen, chenshangfeng@mail.iap.ac.cn

Abstract

A recent study revealed an impact of the intensity of early spring Aleutian low (AL) on the succeeding winter ENSO. This study examines the ability of 41 climate models that participated in CMIP6 in simulating the early spring AL–winter ENSO connection. It is shown that there exists a large diversity among the models in simulating this AL–ENSO linkage. A number of models capture well the observed AL–ENSO connection and the associated physical processes. However, the AL–ENSO relation in several models is opposite to the observed. Diversity of the AL–ENSO connection is related to the spread in the spatial structure of AL-related atmospheric anomalies over the North Pacific. In the models that capture the observed AL–ENSO connection, weakened AL induces an anomalous anticyclone over the northern middle and high latitudes and an anomalous cyclone over the subtropical North Pacific. The resultant westerly wind anomalies over the tropical western-central Pacific (TWCP) induce an El Niño sea surface temperature (SST) anomaly pattern in the following winter. By contrast, in the models with the AL–ENSO relation opposite to the observations, the AL-associated anomalous anticyclone over the North Pacific extends too southward. As such, the subtropical North Pacific is dominated by northeasterly wind anomalies and SST cooling. The subtropical North Pacific SST cooling induces easterly wind anomalies over the TWCP via wind–evaporation–SST feedback, and leads to a La Niña anomaly pattern in the following winter. The spread in the spatial structure of the AL-associated atmospheric anomalies over the North Pacific is partly due to the diversity in the amplitude of the climatological mean flow.

Significance Statement

A recent study suggested that variation of the AL intensity in early spring could exert a significant impact on the following winter ENSO. It indicated that inclusion of the early spring AL signal could improve the prediction of ENSO and to some extent help reduce the spring predictability barrier of ENSO. To employ the AL as a predictor in the ENSO prediction and forecast, the current climate model should have the ability to simulate realistically the early spring AL variation as well as the physical process linking the early spring AL with the subsequent winter ENSO. Hence, this study examines the performance of the current coupled climate models that participated in the phase 6 of the Coupled Model Intercomparison Project (CMIP6) in simulating the linkage between the early spring AL and the following winter ENSO. We show that there exists a large diversity among the CMIP6 models in simulating the early spring AL–winter ENSO connection. A number of models capture well the observed AL–ENSO connection and the associated physical processes. However, the AL–ENSO relation in several models is opposite to the observed. The factors leading to the spread are further examined. Results of this study would have implications in improving our understanding of the impact of extratropical atmospheric forcing on the ENSO and improving the seasonal forecasting of the ENSO.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Shangfeng Chen, chenshangfeng@mail.iap.ac.cn

1. Introduction

El Niño–Southern Oscillation (ENSO) is the strongest air–sea coupling system in the tropics, featured by large sea surface temperature (SST) anomalies in the tropical central-eastern Pacific (Bjerknes 1969; Philander 1990; McPhaden et al. 2006; Alexander et al. 2002). The occurrence of ENSO has pronounced impacts on the weather, climate, agriculture, and hydrological cycle over many parts of the world (e.g., Zhou and Chan 2007; Cai et al. 2011; Cheung et al. 2012; Zhang et al. 2012, 2016, 2017; Luo et al. 2016; Jia et al. 2016; Yeh et al. 2015, 2018a; Domeisen et al. 2019; Wang 2019). Due to the critical impact of ENSO, it is of great importance to examine the factors leading to the occurrence of ENSO and to improve the prediction skill of ENSO. It is now well recognized that the processes within the tropical Pacific play an important role in the evolution and phase transitions of ENSO (Schopf and Suarez 1988; Battisti 1988; Philander 1990; Jin 1997, 1999; Weisberg and Wang 1997; Ren et al. 2016; Tang et al. 2018). Particularly, surface zonal wind anomalies over the tropical western-central Pacific (TWCP) induce SST anomalies in the tropical central-eastern Pacific via triggering eastward-propagating Kelvin waves (Lengaigne et al. 2004; Anderson and Perez 2015; Ren et al. 2016). The generated SST anomalies in the tropical central-eastern Pacific further develop to an ENSO event via positive air–sea interaction.

Although the understanding of the ENSO theory and the performance of climate models in simulating ENSO have been much improved in the past several decades (Luo et al. 2005; Tang et al. 2018; Zhang et al. 2020), the prediction skill of ENSO in the current dynamical climate models has declined significantly since the early 2000s (Barnston et al. 2012; Tang et al. 2018). The decrease in ENSO predictability since the early 2000s is suggested to be associated with the increased occurrence of the central Pacific El Niño events (Ashok et al. 2007; Yu and Kim 2011). Studies indicated that atmosphere–ocean forcing outside the tropical Pacific plays a critical role in the increased occurrence of the central Pacific ENSO (Yu et al. 2012; Di Lorenzo et al. 2015). Therefore, in addition to processes within the tropical Pacific, recent studies have demonstrated that the climate system outside the tropical Pacific contributes significantly to the occurrence and development of ENSO, including Arctic sea ice anomalies (Chen et al. 2020b; Kim et al. 2020; Heo et al. 2021), SST anomalies in the Indian and Atlantic Oceans (Luo et al. 2010; Wang et al. 2011; Ding et al. 2012; Ham et al. 2013; Ham and Kug 2015), SST anomalies in the North Pacific (Chang et al. 2007; Wang et al. 2012; Ding et al. 2017; Su et al. 2018; Zhang et al. 2021), and extratropical atmospheric variability (Vimont et al. 2001; Nakamura et al. 2006; Chen et al. 2014; Ding et al. 2015; Yeh et al. 2018b; Min and Zhang 2020).

Previous studies indicated that ENSO influences the Aleutian low (AL) (Horel and Wallace 1981; Straus and Shukla 2000; Wang et al. 2000; Alexander et al. 2002; Zhang et al. 2022; Larson et al. 2022). A recent study suggested that variation of the AL intensity in early spring (March) could also exert a significant impact on the occurrence of ENSO (Chen et al. 2020a). The AL variation can be defined as the first empirical orthogonal function (EOF) mode of atmospheric variability over the North Pacific (Yu and Kim 2011; Song and Duan 2015). Chen et al. (2020a) reported that when the March AL is weaker than normal, a significant cyclonic anomaly could be induced over the subtropical North Pacific via the wave–mean flow interaction. The associated low-level westerly wind anomalies over the TWCP to the south flank of the anomalous cyclone subsequently impact the following winter El Niño. The reverse is true for the impact of a strengthened AL in March on the following winter La Niña. The impact of the March AL on the following winter ENSO is also found to be independent of ENSO cycle, and the prediction skill of ENSO is significantly improved by considering the preceding AL signal in an empirical model (Chen et al. 2020a). Overall, inclusion of the early spring AL signal could improve the prediction of ENSO. However, in order to employ the AL as a predictor in the ENSO prediction and forecast, the current climate model should have the ability to simulate realistically the early spring AL variation as well as the physical process linking the early spring AL with the subsequent winter ENSO. Therefore, one of the main goals of this study is to examine the performance of the current coupled climate models that participated in phase 6 of the Coupled Model Intercomparison Project (CMIP6; Eyring et al. 2016) in simulating the linkage between the March AL and the following winter ENSO. Our results show that there exists a large diversity among the CMIP6 models in simulating the AL–ENSO connection. In particular, a number of models capture well the observed AL–ENSO connection, whereas the relationship is opposite to the observed in several others. The present study compares the two types of models to explore the factors responsible for the spread of the AL–ENSO relations in the models.

The rest of this study is organized as follows. Section 2 describes the data and methods. Section 3 shows the relationship between the March AL and the subsequent winter ENSO in the observations. Section 4 evaluates the ability of the CMIP6 models in simulating the AL–ENSO connection. Section 5 examines the factors responsible for the large diversity of the AL–ENSO relation. Section 6 provides a summary and discussion.

2. Data and methodology

a. Observational and reanalysis datasets

This study employs monthly sea level pressure (SLP), geopotential height and winds from three different reanalysis datasets, including the National Centers for Environmental Prediction–Department of Energy (NCEP–DOE) Reanalysis from January 1979 to the present (Kanamitsu et al. 2002), the Japanese 55-year Reanalysis (JRA-55) from January 1958 to the present (Ebita et al. 2011), and the fifth-generation European Centre for Medium-Range Weather Forecasts reanalysis (ERA5) from January 1979 to the present (Hersbach and Dee 2016). Atmospheric variables from NCEP–DOE, JRA-55, and ERA5 have horizontal resolutions of 2.5° × 2.5°, 1.25° × 1.25°, and 1° × 1°, respectively.

Monthly SST data are extracted from the National Oceanic and Atmospheric Administration (NOAA) Extended Reconstructed SST version 5 dataset (ERSSTv5) (Huang et al. 2017). The SST data have a horizontal resolution of 2° × 2° in the latitude–longitude grid and span from January 1854 to the present (Huang et al. 2017). We use monthly precipitation data from the Global Precipitation Climatology Project version 2.3 (GPCP), which have a horizontal resolution of 2° × 2° in the latitude–longitude grids and are available from January 1979 to the present (Adler et al. 2003). To facilitate comparison and description, the atmospheric data from the three atmospheric reanalyses, SST data from ERSSTv5, and precipitation data from GPCP are referred to as “observational data.”

b. Model outputs

This study uses monthly outputs of historical simulations of 41 coupled global climate models that participated in CMIP6, which are available from January 1850 to December 2014 (Eyring et al. 2016). Detailed descriptions, including model horizontal resolutions of the atmospheric component, institutions, and model abbreviations, are presented in Table 1. The historical simulations of the CMIP6 models were forced by historical anthropogenic forcing (time-varying land use, ozone, greenhouse gas, anthropogenic aerosol, etc.) and natural forcing (including volcanic eruption, solar radiation, etc.) (Eyring et al. 2016). A number of models only have one realization, and we only employ the first run in each model for a fair comparison.

Table 1

Information of the 41 CMIP6 models employed in this study.

Table 1

c. Methodology

The analysis period is 1979–2014 during which all the observed and model data are available. All the data were computed by removing seasonal cycles relative to the monthly mean climatology over the study period. To calculate the pattern correlation between the observed and simulated fields and calculate the ensemble mean of the simulated results, all the data are converted to a common horizontal resolution of 2° × 2°. The multimodel ensemble mean is calculated as the equal-weighted mean of individual models. This study examines the interannual relation of AL variation with the ENSO, and thus all the variables are subjected to a 9-yr high-pass Lanczos filter to obtain their interannual components (Duchon 1979). Using a 7- or an 11-yr high-pass filter leads to similar results. Notice that the high-pass Lanczos filter can well remove the long-term trend, but may be not able to remove the natural climate variability. This study uses the Niño-3.4 SST index, defined as the area-mean SST anomalies over 5°S–5°N, 120°–170°W, to represent the ENSO variability. It is well known that ENSO has a prominent quasi-biennial cycle and winter ENSO has a significant impact on the winter–spring AL variation (Wang et al. 2000; Chen et al. 2014; Zhang et al. 2017). To ensure that the impact of the March AL on the following winter [D(0)JF(1)] ENSO is not due to the ENSO cycle, ENSO signals have been removed from all the variables in all months by means of linear regression with respect to the Niño-3.4 SST index in preceding winter [D(−1)JF(0)] following Chen et al. (2020a), which is described as follows:
Vres= VNINO34×R.
Here, NINO34 is the D(−1)JF(0) Niño-3.4 SST index, which is usually used to represent the ENSO variability; R is the regression coefficient of the variable V onto the D(−1)JF(0) Niño-3.4 SST index; and Vres is the part of the variable V that the winter ENSO signal has been linearly removed.

3. Relation of early spring AL with the following ENSO in the observations

We first use an EOF technique to extract the leading patterns of interannual SLP variations over the North Pacific. The domain for the EOF analysis spans from 20° to 70°N and from 120°E to 100°W. A slight change in the domain in the EOF analysis does not alter the results. Anomalous SLP fields have been weighted by the cosine of latitude to take into account of the decrease of area with the latitude increase prior to the EOF analysis (North et al. 1982a). Figure 1 shows the EOF1 and EOF2 of monthly SLP anomalies over the North Pacific. Similar EOF patterns are obtained from the NCEP–DOE, ERA5, and JRA-55 (Fig. 1). In the following, we mainly describe the results derived from the NCEP–DOE unless otherwise stated. EOF1 and EOF2 explain about 29% and 21% of the total variance, respectively. Following the method of North et al. (1982b), EOF1 and EOF2 can be separated from each other. The spatial pattern of EOF1 features same-sign SLP anomalies over the North Pacific corresponding to the region where climatological AL is located, with a maximum center of action around 50°N, 165°W (Figs. 1a,c,e). Therefore, EOF1 of SLP anomalies represents variation of the AL intensity, consistent with previous studies (Yu and Kim 2011; Song and Duan 2015; Chen et al. 2020a). According to the pattern of EOF1, we define an AL intensity (ALI) as regional-mean SLP anomalies over 30°–65°N, 160°E–160°W to describe interannual variation of the AL intensity. The correlation coefficient between the ALI and the principal component (PC) of EOF1 is as high as 0.99. EOF2 is characterized by a dipolar SLP anomaly, representing the North Pacific Oscillation (NPO) with two centers of action located around 35°N, 165°W and 60°N, 180°, respectively (Yu and Kim 2011; Chen et al. 2020a).

Fig. 1.
Fig. 1.

The (a),(c),(e) first and (b),(d),(f) second EOF modes of monthly SLP anomalies over the North Pacific (20°–70°N, 120°E–100°W) during 1979–2014. SLP data are obtained from (a),(b) NCEP–DOE, (c),(d) ERA5, and (e),(f) JRA-55. The intensity index of the Aleutian low (ALI) is defined as area-mean SLP anomalies over the box region shown in (a).

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Before evaluating performance of the CMIP6 models in simulating the AL–ENSO relation, we first examine the linkage between the early spring (March) AL and the following winter ENSO in the observations. The observed correlation coefficient between the March [denoted as Mar(0) for short] ALI and the following winter [denoted as D(0)JF(1) for short] Niño-3.4 SST index reaches 0.38 during 1979–2014, significant at the 5% level, consistent with Chen et al. (2020a). Here and in the remainder of this study, the time notations “−1,” “0,” and “1” indicate the year before, during, and after the March AL year. Notice that the AL intensity is below (above) normal when ALI is larger (less) than zero, with positive (negative) SLP anomalies over mid and high-latitude North Pacific.

Figure 2 displays evolutions of SST, precipitation and 850-hPa wind anomalies from Mar(0) to the following D(0)JF(1) obtained by regression upon the normalized Mar(0) ALI. When Mar(0) AL is below normal, an apparent dipolar atmospheric anomaly pattern appears over the North Pacific, with a marked anticyclonic anomaly over midlatitudes and a pronounced cyclonic anomaly over the subtropical northwestern Pacific (Fig. 2b). Decrease in the AL intensity and associated easterly wind anomalies weaken the westerly jet over the midlatitude North Pacific. Weakening of the westerly jet leads to decrease in storm track activity, which is accompanied by cyclonic vorticity forcing to its south, explaining the formation of the cyclonic anomaly over the subtropical North Pacific (Lau 1988; Chen et al. 2014, 2020a). The westerly wind anomalies to the south of the cyclonic anomaly further impact the following winter El Niño via triggering eastward-propagating and downwelling Kelvin waves (Huang et al. 2001; Lengaigne et al. 2004; Nakamura et al. 2006; Chen et al. 2014; Wang et al. 2019a). The reverse conditions are true for the impact of the strengthened AL in March on the La Niña in the subsequent winter.

Fig. 2.
Fig. 2.

Regression maps of SST anomalies (unit: °C) in (a) Mar(0), (c) AM(0), (e) JJA(0), (g) SON(0), and (i) D(0)JF(1) upon the normalized ALI in Mar(0). Regression maps of precipitation (shading; unit: mm day−1) and 850-hPa winds (vectors; m s−1) anomalies in (b) Mar(0), (d) AM(0), (f) JJA(0), (h) SON(0), and (j) D(0)JF(1) upon the normalized ALI in Mar(0). Both directions of wind anomalies that do not exceed the 95% confidence level are not shown. Stippled regions indicate (left) SST and (right) precipitation anomalies significant at the 5% level.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Significant decreases in precipitation are seen around 30°–50°N extending from the regions off the west coast of North America to central Pacific. These negative precipitation anomalies form due to downward motion anomalies related to the anticyclonic anomaly and due to northeasterly wind anomalies that bring colder and drier air from higher latitudes (Fig. 2b). In addition, significant increases in precipitation are seen over the subtropical North Pacific, which is attributed to the anomalous low-level convergence induced by the cyclonic anomaly as well as attributed to the anomalous southwesterly winds that bring moister air from lower latitudes (Fig. 2b). SST anomalies in Mar(0) are very weak in the tropical Pacific (Fig. 2a). Note that Mar(0) ALI-related SST anomalies in preceding winter in the tropical Pacific are weak (not shown). This confirms that the early spring AL–winter ENSO relation is independent of the ENSO cycle. SST anomalies in the extratropical North Pacific show a horseshoe-like pattern, with significant cooling along the west coast of Canada and to the east of the Hawaiian Islands, and prominent warming extending from south of Japan eastward to central Pacific (Fig. 2a). Formation of Mar(0) SST anomalies can be explained by atmospheric anomalies related to the Mar(0) ALI (Chen et al. 2020a). For example, the northerly wind anomalies along west coast of Canada carry cold and dry air from higher latitudes and lead to SST cooling there (Fig. 2a). In addition, the anticyclonic anomaly over the midlatitude North Pacific is accompanied by decrease in total cloud cover, leads to increase in the downward shortwave radiation (not shown) and contributes to SST warming.

In AM(0), the dipolar atmospheric anomaly pattern is maintained over the North Pacific, but the anticyclonic anomaly over the midlatitudes weakens (Fig. 2d). SST anomalies in the extratropical North Pacific increase (Fig. 2c). In addition, significant SST warming is apparent over the tropical central Pacific with a northeastward extension. Studies have indicated that SST warming in the subtropical northeastern Pacific is an important precursor for the El Niño occurrence in the following winter (Chiang and Vimont 2004; Chang et al. 2007; Amaya 2019; Zheng et al. 2021a). The SST anomalies in the subtropical northeastern Pacific can extend to the tropical central Pacific via the wind–evaporation–SST feedback and the summer deep convection mechanism (Xie and Philander 1994; Chang et al. 2007; Amaya 2019; Amaya et al. 2019). In particular, the SST warming in the subtropical North Pacific induced by the cyclonic anomaly related to the Mar(0) AL induces local anomalous atmospheric convection (indicated by positive precipitation anomalies), which in turn helps maintain the anomalous cyclone over the subtropical North Pacific via a Gill-type atmospheric response (Gill 1980; Chen et al. 2020a). In JJA(0), significant SST warming and positive precipitation anomalies are found over the tropical central Pacific (Figs. 2e,f). In addition, westerly wind anomalies are seen over the tropical western-central Pacific (Fig. 2f). Via the Bjerknes positive air–sea feedback (Bjerknes 1969), SST warming, enhanced convection, and westerly wind anomalies maintain and develop to an El Niño event in following winter (Figs. 2e–j).

4. The AL–ENSO relationship in CMIP6 models

In this section, we first examine the ability of the CMIP6 models in simulating spatial patterns of the AL and ENSO. Then, we evaluate the performance of the models in capturing the relationship of the March AL with the following ENSO.

Similar to that in the observations, ALI in the models is defined as regional-mean SLP anomalies over 30°–65°N, 160°E–160°W. Figure 3 shows the spatial distribution of the March AL variation, represented by regression of March SLP anomalies upon the normalized ALI in simultaneous March. In general, all the models simulate well the distribution of SLP anomalies over the North Pacific related to the March AL (Fig. 3), similar to that in the observations (upper left panel of Fig. 3). The pattern correlations of SLP anomalies related to the ALI over the North Pacific (region in Fig. 3) between the observed and simulated are all over 0.8 (Fig. 4a). Compared to the observations, positive SLP anomalies over the mid- and high-latitude North Pacific extend more southeastward in several models, including CanESM5-CanOE, MIROC-ES2L, MIROC6, MRI-ESM2-0, and UKESM1-0-LL. Moreover, the simulated amplitude of the AL is larger in most models compared to the observations (Fig. 4b). Here, the amplitude of the AL is defined as the maximum values of SLP anomalies over the North Pacific in Fig. 3. We note that the observed amplitude of AL is within the range among the 41 CMIP6 models. We then examine the ability of the models in simulating the spatial pattern of ENSO. The spatial pattern of ENSO is represented by SST anomalies in winter regressed upon simultaneous winter Niño-3.4 SST index as ENSO tends to peak in winter. Figure 4c shows pattern correlations between the observed and simulated SST anomalies in the tropical Pacific (10°S–10°N, 120°–9°W). The correlations are larger than 0.8, except for MCM-UA-1-0.

Fig. 3.
Fig. 3.

SLP anomalies in Mar(0) regressed upon the normalized Mar(0) ALI in the observations and historical outputs of 41 CMIP6 models. Stippled regions indicate SLP anomalies significant at the 5% level.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Fig. 4.
Fig. 4.

(a) Pattern correlations of Mar(0) ALI-related simultaneous SLP anomalies over the North Pacific (20°–60°N, 120°E–100°W) between observations and CMIP6 models. (b) Amplitudes of the Mar(0) AL intensity obtained from the observations and CMIP6 models. (c) Pattern correlations of winter Niño-3.4 SST-related simultaneous SST anomalies over tropical Pacific (5°S–5°N, 120°E–60°W) between observations and CMIP6 models. The yellow bar in (a)–(c) represents the ensemble mean of the 41 CMIP6 models, with the error bar indicating one standard deviation of the results among the involved CMIP6 models. Red and blue bars in (b) indicate the ensemble mean of the PC and NC CMIP6 models, respectively. Definitions of the PC and NC CMIP6 models are given in the main text. The green bar in (b) indicates the observed result.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Next, we examine the simulation of the early spring AL–winter ENSO relation. Figure 5 shows the SST anomalies in D(0)JF(1) obtained by regression upon the ALI in preceding March. There is a large spread among the models. Particularly, SST anomalies in D(0)JF(1) in the tropical Pacific are weak in majority of models (Fig. 5). Several models, including CESM2-WACCM, CNRM-CM6-1-HR, FGOALS-f3-L, FGOALS-g3, MPI-ESM1-2-LR, and NorESM2-LM, capture well the observed AL–ENSO connection. In these models, March ALI has a significant positive correlation with the following winter Niño-3.4 SST index (Fig. 7). This suggests that, in these models, the decrease in the March ALI is accompanied by an El Niño pattern in the following winter. By contrast, the decrease in March ALI is associated with a La Niña anomaly pattern in the subsequent winter in CanESM5-CanOE, GISS-E2-1-G, MIROC-ES2L, and MIROC6 (Fig. 7), which is in sharp contrast with that in the observations.

Fig. 5.
Fig. 5.

SST anomalies (unit: °C) in D(0)JF(1) regressed upon the normalized ALI in preceding Mar(0) for the period of 1979–2014 in the 41 CMIP6 models. Stippled regions indicate SST anomalies significant at the 5% level.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

ENSO-related SST anomalies in the tropical Pacific affect significantly the overlying atmosphere circulation (Bjerknes 1969). The diversity of the models in simulating the AL–ENSO relation is reflected in the spread of the Mar(0) ALI-related D(0)JF(1) geopotential height anomalies (Fig. 6). Corresponding to the SST warming in the tropical central-eastern Pacific in CESM-WACCM, CNRM-CM6-1-HR, FGOALS-f3-L, FGOALS-g3, MPI-ESM1-2-LR, and NorESM2-LM, large-scale positive geopotential height anomalies at 200 hPa are seen over the tropical Pacific in D(0)JF(1) as a Gill-type atmospheric response (Fig. 6). By contrast, reverse conditions are found in CanESM5-CanOE, GISS-E2-1-G, MIROC-ES2L, and MIROC6 (Fig. 6).

Fig. 6.
Fig. 6.

Geopotential height anomalies at 200 hPa (unit: m) in D(0)JF(1) regressed upon the normalized ALI in preceding Mar(0). Stippled regions indicate anomalies significant at the 5% level.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

5. Factors contributing to the diversity of simulated AL–ENSO relations

The above analyses indicate that current climate models have a large diversity in simulating the early spring AL–winter ENSO relation. In the following, two groups of models are chosen for comparison to reveal the factors leading to the spread in the AL–ENSO relation. According to Fig. 7, CESM-WACCM, CNRM-CM6-1-HR, FGOALS-f3-L, FGOALS-g3, MPI-ESM1-2-LR, and NorESM2-LM are chosen as positive correlation group models (denoted as PC) as the AL–ENSO relation in these models is similar to the observed (Fig. 7). The ensemble mean of the correlation coefficient between the Mar(0) AL and D(0)JF(1) Niño-3.4 index in the PC group is 0.42 (red bar in Fig. 7). In contrast, CanESM5-CanOE, GISS-E2-1-G, MIROC-ES2L, and MIROC6 are selected as negative correlation (NC) group models. In the NC group, the AL–ENSO relation is opposite to the observed (Fig. 7). The ensemble mean of the Mar(0) AL-D(0)JF(1) Niño-3.4 correlation in the NC group is −0.4 (blue bar in Fig. 7).

Fig. 7.
Fig. 7.

Correlation coefficients of the Mar(0) ALI with the D(0)JF(1) Niño-3.4 SST index in the observation and 41 CMIP6 models. Red and blue bars indicate the ensemble mean of the correlation coefficients for the PC and NC group models, with error bars indicating one standard deviation among the involved models. A green bar indicates the observed result.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Figure 8 displays evolutions of ensemble mean SST anomalies from Mar(0) to D(0)JF(1) for the PC and NC groups as well as their differences. The evolution of SST anomalies in the Pacific for the PC group (Figs. 8a,d,g,j,m) is similar to that in the observations (Figs. 2a,c,e,g,i). In Mar(0), SST anomalies are weak in the tropical Pacific. A horseshoe-like SST anomaly pattern appears in the extratropical North Pacific. SST warming starts to occur in the tropical central-eastern Pacific in AM(0) (Fig. 8d) and develops to an El Niño anomaly pattern in the following winter (Fig. 8m). The evolution of SST anomalies in the NC group shows notable differences from that in the PC group. In Mar(0), a horseshoe-like SST anomaly pattern in the extratropical North Pacific displays a southwestward extension in the NC group (Fig. 8b). In particular, significant SST cooling is seen to extend southwestward from the west coast of North America to the tropical central Pacific (Chiang and Vimont 2004; Chang et al. 2007; Vimont et al. 2009; Amaya 2019). SST cooling is seen in the tropical central-eastern Pacific in JJA(0) (Fig. 8h), and then develops to a La Niña pattern in the following winter (Fig. 8n). In the difference maps of the Mar(0) ALI-related SST anomalies between the PC and NC in Mar(0), a tripole SST difference pattern is seen in the North Pacific, with marked SST warming differences in the region extending northeastward from the tropical central Pacific and in the high-latitude North Pacific, together with pronounced SST cooling differences in the subtropical western North Pacific (Fig. 8c). Formation of the tripole SST difference pattern in Mar(0) is related to the difference of atmospheric circulation anomalies as will be described below. From AM(0) to D(0)JF(1), SST anomalies in the tropical central-eastern Pacific are significantly larger in the PC group than in the NC group, showing a clear El Niño difference pattern (Figs. 8f,i,l,o).

Fig. 8.
Fig. 8.

SST anomalies (unit: °C) in (a),(b) Mar(0) ,(d),(e) AM(0), (g),(h) JJA(0), (j),(k) SON(0), and (m),(n) D(0)JF(1) regressed upon the Mar(0) ALI for the ensemble means of the (left) PC and (center) NC CMIP6 models. (c),(f),(i),(l),(o) The SST differences between the ensemble mean of the PC and NC models. Stippled regions in the left and center columns indicate SST anomalies significant at the 5% level. (right) Stippled regions indicate SST differences that are significantly different from zero at the 5% level. The box in (c) covers the region of 2.5°–17.5°N, 155°E–130°W.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Evolutions of precipitation anomalies (Fig. 9) correspond well to the evolutions of SST anomalies (Fig. 8). In the PC group in Mar(0), positive precipitation anomalies are seen over the subtropical western-central North Pacific and negative precipitation anomalies are found along the west coast of Canada with a westward extension (Fig. 9a), bearing a close resemblance to those in the observations (Fig. 2b). In AM(0), precipitation anomalies over the North Pacific are mostly insignificant. In JJA(0), significant positive precipitation anomalies are seen in the tropical central-eastern Pacific (Fig. 9g). Those anomalies are maintained with enhancement in the following SON(0) and D(0)JF(1) (Figs. 9j,m), consistent with the development of SST warming there (Figs. 8g,j,m). In the NC group, the precipitation anomaly pattern over the extratropical North Pacific is similar to that in the PC group. However, the tropical western North Pacific is covered by large negative precipitation anomalies in the NC group (Fig. 9b), which can be clearly detected in the difference map (Fig. 9c). The large negative precipitation anomalies in the tropical western North Pacific induce strong easterly wind anomalies over the tropical equatorial western Pacific via a Gill-type atmospheric response (Huang et al. 2001; Lian et al. 2014; Lai et al. 2015), which contribute to formation of a La Niña anomaly pattern in the following winter. Negative precipitation anomalies are seen in the tropical central Pacific in AM(0) (Fig. 9e), extend eastward to the tropical eastern Pacific in JJA(0) (Fig. 9h), and are maintained and enhanced in the following winter (Fig. 9n). In addition, as expected, differences of the precipitation anomalies between PC and NC show that Mar(0) ALI-related precipitation anomalies are significantly larger in the tropical central-eastern Pacific in the PC compared to those of NC from JJA(0) to D(0)JF(1) (Figs. 9i,l,o), consistent with the SST differences (Figs. 8i,l,o).

Fig. 9.
Fig. 9.

As in Fig. 8, but for precipitation anomalies (unit: mm day−1).

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

The above analyses indicate that one of the most prominent differences between the PC and NC groups is the spring SST anomalies in the subtropical North Pacific (SNP) (Figs. 8c,f). In the NC group, significant negative SST anomalies appear in the SNP in spring (Figs. 8b,e), which is in sharp contrast to those in the observations (Figs. 2a,c) and in the PC group (Figs. 8a,d). Studies have demonstrated that SST cooling (warming) in the SNP in boreal spring could lead to a La Niña (an El Niño) SST pattern in the following winter (Chiang and Vimont 2004; Chang et al. 2007; Alexander et al. 2010; Amaya 2019; Zheng et al. 2021a). To further confirm the role of the spring SST anomalies in the SNP in modulating evolution of SST anomalies in the tropical Pacific, we defined a SNP SST index as area-average SST anomalies over 2.5°–17.5°N, 155°E–130°W (box in Fig. 8c). Figure 10 displays evolutions of SST anomalies from spring to winter obtained by regression upon the inverted spring SNP SST index in the observations and for the ensemble mean of the NC models. It indicates that significant negative SST anomalies in the SNP in boreal spring can evolve into a La Niña anomaly pattern in the succeeding winter. In particular, a tripolar SST anomaly pattern occurs in the North Pacific in boreal spring in association with SST decrease in the SNP (Figs. 10a,b). Spring negative SST anomalies in the SNP extend to the tropical central Pacific in the following summer via wind–evaporation–SST feedback (Chiang and Vimont 2004; Chang et al. 2007; Alexander et al. 2010; Amaya 2019; Zheng et al. 2021a). The SST cooling in the tropical Pacific appears in summer and develops to a La Niña anomaly pattern in the following winter via positive air–sea interaction (Bjerknes 1969; Chang et al. 2007; Chen et al. 2014, 2020a; Amaya 2019; Zheng et al. 2021b). The results confirm that spring negative (positive) SST anomalies in the SNP could impact the occurrence of a La Niña (an El Niño) event in the following winter. Hence, the diversity in the AL–ENSO relation among the models could be attributed to the spread in the Mar(0) AL-generated SST anomalies in the SNP.

Fig. 10.
Fig. 10.

Anomalies of SST (unit: °C) in (a) MAM(0), (c) JJA(0), (e) SON(0), and (g) D(0)JF(1) regressed upon the inverted normalized spring SNP SST index in the observation. (b),(d),(f),(h) As in (a), (c), (e), and (g), but for the ensemble mean of the NC CMIP6 models. Stippled regions indicate SST anomalies significant at the 5% level. The SNP SST index is defined as area-average SST anomalies over 2.5°–17.5°N, 155°E–130°W (box in Fig. 8c).

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

A relevant question is why significant SST cooling appears in the subtropical North Pacific in the NC group. Chen et al. (2020a) has demonstrated that the formation of SST anomalies in the North Pacific is attributable to surface heat flux changes induced by atmospheric anomalies related to the Mar(0) AL. In particular, in simultaneous spring, the anomalous northerly winds to the east flank of the anomalous anticyclone over the North Pacific bring cold and dry air from higher latitudes (Fig. 2b), which lead to increase in the upward latent and sensible heat fluxes (Chen et al. 2020a), inducing SST cooling along the west coast of North America (Fig. 2a). In addition, anomalous low-level easterly wind anomalies over the midlatitude North Pacific oppose climatological westerly winds (Fig. 2b), lead to decrease in the total wind speed and contribute to SST warming there (Fig. 2a). The southwesterly wind anomalies over the subtropical North Pacific contribute to SST warming in the subtropical North Pacific in the following April–May [AM(0)] via reduction in total wind speed (Fig. 2c).

We further compare atmospheric anomalies related to the Mar(0) ALI between PC and NC groups. Figures 11a and 11b display ensemble mean SLP and 850-hPa wind anomalies in Mar(0) regressed upon the Mar(0) ALI for the PC and NC group, respectively. A cyclonic anomaly is seen in the PC over the subtropical North Pacific (Fig. 11a). As demonstrated by previous studies, the anomalous westerly winds over the tropical western Pacific to the south of the anomalous cyclone in the PC play an important role in the formation of SST warming in the tropical central-eastern Pacific via triggering eastward-propagating and downwelling Kelvin waves (Li 1990; Huang et al. 2001; Nakamura et al. 2006; Chen et al. 2014, 2020b). By contrast, in the NC group, the anomalous anticyclone over the mid- and high-latitude North Pacific extends more southward, as can be clearly indicated by zero value of SLP anomalies (Fig. 11b) and the difference map (Fig. 11c). In particular, the subtropical North Pacific is covered by positive SLP and northeasterly wind anomalies in the NC group (Fig. 11b), in sharp contrast with that in the observations and in the PC group (Figs. 2b and 8a). As such, the anomalous northeasterly wind anomalies enhance surface wind speed and evaporation, leading to SST cooling in the subtropical North Pacific. The subtropical SST cooling further contributes to the formation of a La Niña anomaly pattern in the subsequent winter via the wind–evaporation–SST feedback and tropical air–sea interaction (Chiang and Vimont 2004; Chang et al. 2007; Vimont et al. 2009; Amaya 2019; Zheng et al. 2021a). Therefore, the large difference in the early spring SST anomalies in the subtropical North Pacific between the PC and NC is due to spread of the atmospheric circulation anomalies over the North Pacific.

Fig. 11.
Fig. 11.

SLP (shading; unit: hPa) and 850-hPa winds (vectors; unit: m s−1) anomalies in Mar(0) regressed upon the Mar(0) ALI for the ensemble means of the (a) PC and (b) NC CMIP6 models. (c) The differences between the ensemble means of the PC and NC models. Stippled regions in (a) and(b) indicate SLP anomalies significant at the 5% level. Stippled regions in (c) indicate the SLP differences that are significantly different from zero at the 5% level. Wind anomalies in both directions less than 0.2 m s−1 are not shown. The bold black line indicates zero values of SLP anomalies. The box in (c) covers the region of 0°–10°N, 120°E–170°W.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Figure 12a shows a scatterplot of Mar(0) ALI-related Mar(0) SLP anomalies averaged over the SNP against Mar(0) SST anomalies averaged in the SNP. The two variables in Fig. 12a are significantly correlated with each other, with a correlation coefficient of −0.69 among the 41 models. It indicates that a model’s ability in simulating the Mar(0) SLP anomalies over the SNP has a close relation with that in capturing the Mar(0) SST anomalies in the SNP. If a model produces positive (negative) SLP anomalies over the SNP, negative (positive) SST anomalies tend to be induced in the SNP (Fig. 12a) and easterly (westerly) wind anomalies tend to develop over the tropical western-central Pacific (Figs. 12b,d), which contributes to the AL–ENSO relation opposite (similar) to the observed (Fig. 12c).

Fig. 12.
Fig. 12.

(a) Scatterplot of Mar(0) ALI-related Mar(0) SLP anomalies averaged over SNP against Mar(0) SST anomalies averaged in SNP. (b) Scatterplot of Mar(0) ALI-related Mar(0) 850-hPa zonal wind anomalies averaged over the tropical western-central Pacific (TWCP; box in Fig. 11c) against Mar(0) SST anomalies averaged in SNP. (c) Scatterplot of Mar(0) ALI-related Mar(0) 850-hPa zonal wind anomalies averaged over TWCP against the Mar(0) AL–D(0)JF(1) ENSO correlation. (d) Scatterplot of Mar(0) ALI-related Mar(0) SLP anomalies averaged over SNP against Mar(0) 850-hPa zonal wind anomalies averaged over TWCP.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Why there exists a clear cyclonic anomaly over the SNP in the PC group, but is missing in the NC group? Chen et al. (2020a) indicated that the formation of the Mar(0) ALI-related cyclonic anomaly is due to the interaction between the low frequency mean flow and synoptic-scale eddies, as has been described in section 2. Studies indicated that the intensity of the wave–mean flow interaction is closely related to the low-frequency mean flow and the synoptic-scale eddies (Jin et al. 2006a,b; Jin 2010; Chen et al. 2015). If the activity of synoptic-scale eddies is similar in two states, the intensity of the wave–mean interaction may be mainly determined by the low-frequency mean flow (Jin et al. 2006a,b; Jin 2010; Chen et al. 2015). Stronger low-frequency mean flow would lead to a stronger feedback of synoptic-scale eddies and a stronger cyclonic anomaly over the SNP associated with the Mar(0) ALI. Figures 13a and 13b show differences of mean zonal winds at 1000 and 850 hPa between the PC and NC groups. Mean westerly winds to the east of Japan over the midlatitude North Pacific is significantly stronger in the PC group than in the NC group. This confirms the hypothesis that the difference in the low-frequency mean flow may be one of the plausible factors leading to the spread in the atmospheric anomalies over the SNP related to the Mar(0) AL and partly explains the diversity in the AL–ENSO connection.

Fig. 13.
Fig. 13.

Climatology (contours; unit: m s−1) of (a) 1000- and (b) 850-hPa zonal winds for the ensemble mean of the 41 CMIP6 models. Differences (shading; unit: m s−1) in the climatology of (a) 1000- and (b) 850-hPa zonal winds between the ensemble means of the PC and NC models. Stippled regions indicate the differences that are significantly different from zero at the 5% level.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

6. Summary and discussion

A previous study indicated that the early spring AL intensity has a significant impact on the occurrence of the following winter ENSO. When the AL intensity in early spring is weaker (stronger) than normal, an El Niño (La Niña) event tends to occur in the subsequent winter. This study examines the performance of the current climate models that participated in CMIP6 in simulating the observed AL–ENSO relationship. It is found a large diversity exists in the early spring AL–winter ENSO linkage among the CMIP6 historical simulations. A majority of the models produce a weak connection of the early spring AL variation with the following winter ENSO. Six of the 41 models can well capture the observed AL–ENSO relation. In these models, weakened (strengthened) AL intensity in early spring is accompanied by a following winter El Niño (La Niña) event. In contrast, several models produce an AL–ENSO relation opposite to the observations. In these models, enhancement (weakening) of the early spring AL intensity tends to be accompanied by an El Niño (a La Niña) in the following winter. The factor contributing to the diversity of the AL–ENSO relation in the CMIP6 models is further examined, which is schematically summarized in Fig. 14.

Fig. 14.
Fig. 14.

Schematic diagram for the diversity of the CMIP6 models in simulating impact of the early-spring AL on the following winter ENSO. Light-blue shading indicates the climatology of the westerly jet over the midlatitude North Pacific. Arrows indicate wind anomalies. Red and blue shadings indicate SST warming and cooling, respectively.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

The spread in the early spring AL–winter ENSO relation among the models could be attributed to the diversity in the spatial structure of the March AL. In the PC group, the Mar(0) AL-related atmospheric circulation anomalies over the North Pacific are similar to those in the observations, with a strong anticyclonic anomaly over the mid- and high-latitude North Pacific and a significant cyclonic anomaly over the subtropical North Pacific when Mar(0) AL is weaker than normal (Fig. 14a). The westerly wind anomalies over the tropical western-central Pacific lead to SST warming in the tropical central-eastern Pacific. The summertime SST warming in the tropical central-eastern Pacific further develops to an El Niño anomaly pattern in winter via a positive air–sea interaction (Fig. 14a).

By contrast, in the NC group, the weakened Mar(0) AL-related anticyclonic anomaly extends too southward (Fig. 14b). The SNP is covered by positive SLP and northeasterly wind anomalies, leading to a SST cooling in the subtropical northeastern Pacific. The subtropical northeastern Pacific SST cooling induces strong easterly wind anomalies over the tropical western-central Pacific via the wind–evaporation–SST feedback, which further contribute to the formation of a La Niña SST pattern in the following winter (Fig. 14b).

Further analyses indicate that the difference in the atmospheric anomalies over the subtropical North Pacific related to the Mar(0) AL may be partly due to the difference in climatological mean flow. The westerly jet stream over the midlatitude North Pacific is much stronger in the PC group than that in the NC group. Strong westerly winds lead to a strong wave–mean flow interaction that contributes to the formation of a cyclonic anomaly over the subtropical North Pacific for the PC models.

From Figs. 3 and 4b, there exists a large diversity in the amplitude of the Mar(0) AL among the models. A relevant question is whether the ability of the models to simulate the Mar(0) AL-related atmospheric anomalies over the subtropical North Pacific is related to the Mar(0) AL’s intensity. To address this issue, we have compared the intensity of the Mar(0) AL between the PC and NC groups (Fig. 4b). The ensemble mean Mar(0) AL intensity is weaker in the PC group compared to that in the NC group (Fig. 4b). However, the difference in the amplitude of the ALI between the PC and NC models is not significantly different from zero at the 10% level. This suggests that the diversity in the AL–ENSO relation among the models is not likely related to the spread in the ALI amplitude.

As seen from Table 1, the CNRM-CM6-1-HR and CNRM-CM6-1 come from the same institute, with a higher horizontal resolution in the former. From Fig. 7, the CNRM-CM6-1-HR captures a significant positive correlation between the March AL and the following winter ENSO, similar to the observed. By contrast, the CNRM-CM6-1 produces a weak negative correlation, opposite to the observed. The horizontal resolution is much lower in MPI-ESM-1-2-LR compared to the MPI-ESM-1-2-HR, both of which are from the same climate center. MPI-ESM-1-2-LR produces a significant positive relation of the Mar(0) AL with the following winter ENSO, consistent with the observed. In contrast, the MPI-ESM-1-2-HR produces a weak negative correlation between the Mar(0) AL and winter ENSO. This suggests that the diversity of the Mar(0) AL–winter ENSO relation is not likely related to the difference in the models’ horizontal resolution.

This study indicates that the spread of the early spring AL–winter ENSO relation in the CMIP6 models is due to spread of the spatial structure of atmospheric anomalies over the North Pacific related to the early spring AL. Studies have indicated that winter ENSO could exert impacts on the spatial structure and intensity of the following early spring AL via atmospheric teleconnection (Wang et al. 2000; Alexander et al. 2002). Therefore, to ensure that the close connection of early spring AL to the following winter ENSO is not due to the ENSO cycle, we have removed the winter ENSO signal from the analysis fields via a linear regression method as has been described in section 2. However, one may argue that the linear regression method may be not able to totally remove the ENSO signal. Hence, to examine whether the spread of the early spring AL–winter ENSO linkage is related to the spread of ENSO’s behavior, including its amplitude and evolutions, we have checked standard deviations of the D(−1)JF(0) Niño-3.4 SST index in the CMIP6 models in Fig. 15. The ensemble mean of the standard deviation of D(−1)JF(0) Niño-3.4 SST index is weaker in the PC group compared to that in the NC group (Fig. 15). However, the difference between the PC and NC models is not significant (Fig. 15). In addition, we have compared evolutions of SST anomalies from JJA(−1) to MAM(0) obtained by regression upon the D(−1)JF(0) Niño-3.4 SST index in the PC and NC groups (Fig. 16). It is found that evolutions of SST anomalies in the tropical Pacific are generally similar between the PC and NC groups (Fig. 16). Hence, the spread in the early spring AL–winter ENSO relation among the models is not likely related to the spread in the behaviors of ENSO.

Fig. 15.
Fig. 15.

Standard deviation of winter Niño-3.4 SST index in observations and 41 CMIP6 models. Red (blue) bars indicate the ensemble mean of the PC (NC) group models, with the error bar indicating one standard deviation of the results among the involved models. The green bar indicates the observed result.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

Fig. 16.
Fig. 16.

SST anomalies (unit: °C) in (a),(b) JJA(−1), (d),(e) SON(−1), (g), (h) D(−1)JF(0), and (j),(k) MAM(0) regressed upon the D(−1)JF(0) Niño-3.4 SST index for the ensemble means of the (left) PC and (center) NC CMIP6 models. (c),(f),(i),(l) The SST differences between ensemble means of the PC and NC models. Stippled regions in the left and middle columns indicate SST anomalies significant at the 5% level. (right) Stippled regions indicate the SST differences that are significantly different from zero at the 5% level.

Citation: Journal of Climate 36, 2; 10.1175/JCLI-D-22-0323.1

We have also examined the spatial structure of the SLP and 850-hPa winds anomalies over the North Pacific in association with the early spring AL index when the ENSO signal was not removed (not shown). The results show that if the ENSO signal was not removed, the early spring AL-related anomalous anticyclone over the North Pacific still extends more southward in the NC models compared to the PC models (not shown). In addition, the PC (NC) models still produce a significant positive (negative) correlation of the early spring AL index with the following winter Niño-3.4 SST index. Hence, the results obtained in this study are not sensitive to whether the ENSO signal was removed or not from the analysis data. However, removal of the preceding winter ENSO signal can ensure that the connection of the early spring AO with the following winter ENSO is not due to the ENSO cycle.

The difference in the climatology of the SLP and 850-hPa winds in Fig. 11 is characterized by a dipole structure over the North Pacific. This suggests a northward shift of the early spring AL’s center in the PC models. Previous studies indicated that the climatological mean SST structure in the North Pacific could impact the AL (Mantua and Hare 2002; Wang et al. 2019b). We have examined differences in climatological mean SST in the North Pacific between the PC and NC models (not shown). The differences in climatological mean SST are not significant over most North Pacific (not shown). This implies that the difference in the spatial structure of the early spring AL between the PC and NC models is not likely related to the difference in the climatological mean SST structure in the North Pacific. Studies also indicated that sea ice concentration variation over the North Pacific and Arctic may exert impacts on the North Pacific atmospheric circulation anomalies (Cavalieri and Parkinson 1987; Yeo et al. 2014; Kim et al. 2020). We have examined the difference in the climatological mean sea ice concentration between the PC and NC models (not shown). The differences of the sea ice concentration over the North Pacific are weak (not shown). However, we found that the climatological mean sea ice concentration around the Barents Sea and the Siberian Sea is significantly lower in the PC models (not shown). Whether and how the difference in the sea ice concentration around the Barents Sea and the Siberian Sea contributes to the difference in the spatial structure of the AL between the PC and NC models needs to be further investigated.

Studies have indicated that ENSO events have a prominent winter phase locking phenomenon (Wang et al. 2000; Liu et al. 2021). Specifically, ENSO-related SST, precipitation, and atmospheric anomalies in the tropical Pacific tend to reach their maximum amplitudes in boreal winter. It is noted that not all the CMIP6 models can reproduce the winter phase locking behavior of the ENSO (Liu et al. 2021). An immediate question is whether a model’s ability in capturing the observed relationship between the March AL and the following winter ENSO is related to the model’s performance in reproducing the winter phase locking phenomenon. To address this issue, we calculated the standard deviation of the monthly Niño-3.4 SST index in the 41 CMIP6 models (not shown). We found that all the six PC models (i.e., CESM-WACCM, CNRM-CM6-1-HR, FGOALS-f3-L, FGOALS-g3, MPI-ESM1-2-LR, and NorESM2-LM) can well reproduce the seasonality of the ENSO, with the largest signal in boreal winter, consistent with the observation. For the NC models, except for the GISS-E2-1-G, the CanESM5-CanOE, MIROC-ES2L, and MIROC6 have a good performance in simulating the winter phase locking behavior of the ENSO. Moreover, the winter phase locking phenomenon can be captured by all the other models except for the MCM-UA-1-0. This evidence implies that the diversity of the CMIP6 models in capturing the observed connection of the early spring AL with the following winter ENSO is not likely attributed to the spread of the models in simulating the seasonality of the ENSO.

In observations, the March AL index has a strongest correlation with the following winter [D(0)JF(1)] Niño-3.4 SST index. However, this may not be true in CMIP6 models due to the diverse seasonality of the ENSO. We have computed the lead–lag correlation coefficients between the March ALI with the 3-month-mean Niño-3.4 SST index (not shown). The results show that the PC (NC) models generally exhibit the strongest positive (negative) correlation coefficient between the March ALI and the following winter Niño-3.4 SST index. There indeed exist several PC and NC models that the March ALI is most closely related to the Niño-3.4 SST index in other months [e.g., ND(0)J(1) and JFM(1)], instead of the observed D(0)JF(1). However, their correlation coefficients do not differ significantly from the correlation coefficients between the March ALI and the D(0)JF(1) Niño-3.4 SST index. In addition, we found that the other models (except the PC and NC models) all show weak correlations between the March ALI and the Niño-3.4 SST from previous winter to the following winter. Above evidence ensures that we can use the correlation between the March ALI and the D(0)JF(1) Niño-3.4 SST index uniformly in the model to explore the relationship between the early spring AL and the winter ENSO without affecting our conclusions.

In addition, it should be mentioned that the relationship between the early spring AL and the winter ENSO is unstable. We have calculated the running correlations of the early spring ALI with the following winter Niño-3.4 SST index with different lengths of moving window (not shown). The results show a pronounced enhancement of the impact of the early spring AO on the following winter ENSO around the mid-1990s in the observations. The relative roles of the internal climate variability and global warming in contributing to the recent enhanced impact of the early spring AL on the following winter ENSO would be further investigated. As the connection between the early spring AL and the winter ENSO is unstable, a question is whether the results obtained in this study are sensitive to the selection of the analysis period. To address this issue, we have re-examined the results by using different time periods (not shown). It is found that although the classification of the PC and NC models could be influenced by the selected analysis period to some extent, the main conclusions obtained from this study are robust.

Acknowledgments.

We thank three anonymous reviewers for their constructive suggestions, which helped to improve the paper. This study was supported jointly by the National Natural Science Foundation of China (Grants 42175039, 41961144025, and 41721004), and the Jiangsu Collaborative Innovation Center for Climate Change.

Data availability statement.

The NCEP–DOE Reanalysis data are obtained from http://www.cpc.ncep.noaa.gov/products/wesley/reanalysis2. The JRA-55 data are obtained from https://jra.kishou.go.jp/JRA-55/index_en.html. The ERA5 data are obtained from https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era5. The GPCP precipitation data are obtained from http://www.esrl.noaa.gov/psd/data/gridded/data.gpcp.html. The ERSSTv5 SST data are obtained from https://psl.noaa.gov/data/gridded/data.noaa.ersst.v5.html. The CMIP6 historical outputs are derived from https://esgf-node.llnl.gov/projects/cmip6/.

REFERENCES

  • Adler, R. F., and Coauthors, 2003: The version 2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 11471167, https://doi.org/10.1175/1525-7541(2003)004<1147:TVGPCP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Alexander, M. A., I. Bladé, M. Newman, J. R. Lanzante, N.-C. Lau, and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. J. Climate, 15, 22052231, https://doi.org/10.1175/1520-0442(2002)015<2205:TABTIO>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Alexander, M. A., D. J. Vimont, P. Chang, and J. D. Scott, 2010: The impact of extratropical atmospheric variability on ENSO: Testing the seasonal footprinting mechanism using coupled model experiments. J. Climate, 23, 28852901, https://doi.org/10.1175/2010JCLI3205.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Amaya, D. J., 2019: The Pacific meridional mode and ENSO: A review. Curr. Climate Change Rep., 5, 296307, https://doi.org/10.1007/s40641-019-00142-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Amaya, D. J., Y. Kosaka, W. Zhou, Y. Zhang, S.-P. Xie, and A. J. Miller, 2019: The North Pacific pacemaker effect on historical ENSO and its mechanisms. J. Climate, 32, 76437661, https://doi.org/10.1175/JCLI-D-19-0040.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Anderson, B. T., and R. C. Perez, 2015: ENSO and non-ENSO induced charging and discharging of the equatorial Pacific. Climate Dyn., 45, 23092327, https://doi.org/10.1007/s00382-015-2472-x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ashok, K., S. K. Behera, S. A. Rao, H. Weng, and T. Yamagata, 2007: El Niño Modoki and its possible teleconnection. J. Geophys. Res., 112, C11007, https://doi.org/10.1029/2006JC003798.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Barnston, A. G., M. K. Tippett, and M. L. L’Heureux, 2012: Skill of real-time seasonal ENSO model predictions during 2002–11: Is our capability increasing? Bull. Amer. Meteor. Soc., 93, 631651, https://doi.org/10.1175/BAMS-D-11-00111.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Battisti, D. S., 1988: Dynamics and thermodynamics of a warming event in a coupled tropical atmosphere–ocean model. J. Atmos. Sci., 45, 28892919, https://doi.org/10.1175/1520-0469(1988)045<2889:DATOAW>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97, 163172, https://doi.org/10.1175/1520-0493(1969)097<0163:ATFTEP>2.3.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cai, W., P. van Rensch, T. Cowan, and H. H. Hendon, 2011: Teleconnection pathways of ENSO and the IOD and the mechanisms for impacts on Australian rainfall. J. Climate, 24, 39103923, https://doi.org/10.1175/2011JCLI4129.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cavalieri, D. J., and C. L. Parkinson, 1987: On the relationship between atmospheric circulation and the fluctuations in the sea ice extents of the Bering and Okhotsk Seas. J. Geophys. Res., 92, 71417162, https://doi.org/10.1029/JC092iC07p07141.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chang, P., L. Zhang, R. Saravanan, D. J. Vimont, J. C. H. Chiang, L. Ji, H. Seidel, and M. K. Tippett, 2007: Pacific meridional mode and El Niño–Southern Oscillation. Geophys. Res. Lett., 34, L16608, https://doi.org/10.1029/2007GL030302.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, S. F., B. Yu, and W. Chen, 2014: An analysis on the physical process of the influence of AO on ENSO. Climate Dyn., 42, 973989, https://doi.org/10.1007/s00382-012-1654-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, S. F., B. Yu, and W. Chen, 2015: An interdecadal change in the influence of the spring Arctic Oscillation on the subsequent ENSO around the early 1970s. Climate Dyn., 44, 11091126, https://doi.org/10.1007/s00382-014-2152-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, S. F., W. Chen, R. Wu, B. Yu, and H. Graf, 2020a: Potential impact of preceding Aleutian low variation on the El Niño–Southern Oscillation during the following winter. J. Climate, 33, 30613077, https://doi.org/10.1175/JCLI-D-19-0717.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chen, S. F., R. Wu, W. Chen, and B. Yu, 2020b: Influence of winter Arctic sea ice concentration change on the El Niño–Southern Oscillation in the following winter. Climate Dyn., 54, 741757, https://doi.org/10.1007/s00382-019-05027-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cheung, H. N., W. Zhou, H. Y. Mok, and M. C. Wu, 2012: Relationship between Ural-Siberian blocking and the East Asian winter monsoon in relation to the Arctic Oscillation and the El Niño–Southern Oscillation. J. Climate, 25, 42424257, https://doi.org/10.1175/JCLI-D-11-00225.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chiang, J. C. H., and D. J. Vimont, 2004: Analogous Pacific and Atlantic meridional modes of tropical atmosphere–ocean variability. J. Climate, 17, 41434158, https://doi.org/10.1175/JCLI4953.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Di Lorenzo, E., G. Liguori, N. Schneider, J. C. Furtado, B. T. Anderson, and M. A. Alexander, 2015: ENSO and meridional modes: A null hypothesis for Pacific climate variability. Geophys. Res. Lett., 42, 94409448, https://doi.org/10.1002/2015GL066281.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, H., N. S. Keenlyside, and M. Latif, 2012: Impact of the equatorial Atlantic on the El Niño Southern Oscillation. Climate Dyn., 38, 19651972, https://doi.org/10.1007/s00382-011-1097-y.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, R. Q., J. P. Li, and Y. H. Tseng, 2015: The impact of South Pacific extratropical forcing on ENSO and comparisons with the North Pacific. Climate Dyn., 44, 20172034, https://doi.org/10.1007/s00382-014-2303-5.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ding, R. Q., J. P. Li, Y. H. Tseng, C. Sun, and F. Xie, 2017: Joint impact of North and South Pacific extratropical atmospheric variability on the onset of ENSO events. J. Geophys. Res. Atmos., 122, 279298, https://doi.org/10.1002/2016JD025502.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Domeisen, D. I., C. I. Garfinkel, and A. H. Butler, 2019: The teleconnection of El Niño Southern Oscillation to the stratosphere. Rev. Geophys., 57, 547, https://doi.org/10.1029/2018RG000596.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18, 10161022, https://doi.org/10.1175/1520-0450(1979)018<1016:LFIOAT>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ebita, A., and Coauthors, 2011: The Japanese 55-year reanalysis “JRA-55”: An interim report. SOLA, 7, 149152, https://doi.org/10.2151/sola.2011-038.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eyring, V., S. Bony, G. A. Meehl, C. A. Senior, B. Stevens, R. J. Stouffer, and K. E. Taylor, 2016: Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev., 9, 19371958, https://doi.org/10.5194/gmd-9-1937-2016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat‐induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447462, https://doi.org/10.1002/qj.49710644905.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ham, Y.-G., and J.-S. Kug, 2015: Role of north tropical Atlantic SST on the ENSO simulated using CMIP3 and CMIP5 models. Climate Dyn., 45, 31033117, https://doi.org/10.1007/s00382-015-2527-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ham, Y.-G., J.-S. Kug, J.-Y. Park, and F.-F. Jin, 2013: Sea surface temperature in the north tropical Atlantic as a trigger for El Niño/Southern Oscillation events. Nat. Geosci., 6, 112116, https://doi.org/10.1038/ngeo1686.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Heo, E., S. An, I. Ringgaard, S. Yang, J. Christensen, and M. Sung, 2021: Influence of retreating Barents-Kara sea ice on the periodicity of El Niño–Southern Oscillation. Int. J. Climatol., 42, 48814892, https://doi.org/10.1002/joc.7509.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hersbach, H., and D. Dee, 2016: ERA5 reanalysis is in production. ECMWF Newsletter, No. 147, ECMWF, Reading, United Kingdom, 7, http://www.ecmwf.int/sites/default/files/elibrary/2016/16299-newsletter-no147-spring-2016.pdf.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109, 813829, https://doi.org/10.1175/1520-0493(1981)109<0813:PSAPAW>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, B., and Coauthors, 2017: Extended Reconstructed Sea Surface Temperature version 5 (ERSSTv5), upgrades, validations, and intercomparisons. J. Climate, 30, 81798205, https://doi.org/10.1175/JCLI-D-16-0836.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Huang, R. H., R. H. Zhang, and B. L. Yan, 2001: Dynamical effect of the zonal wind anomalies over the tropical western Pacific on ENSO cycles. Sci. China, 44D, 10891098, https://doi.org/10.1007/BF02906865.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jia, X. J., J. W. Ge, and S. Wang, 2016: Diverse impacts of ENSO on wintertime rainfall over the Maritime Continent. Int. J. Climatol., 36, 33843397, https://doi.org/10.1002/joc.4562.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jin, F. F., 1997: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model. J. Atmos. Sci., 54, 811829, https://doi.org/10.1175/1520-0469(1997)054<0811:AEORPF>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jin, F. F., 1999: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model. J. Atmos. Sci., 54, 811829, https://doi.org/10.1175/1520-0469(1997)054<0811:AEORPF>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jin, F. F., 2010: Eddy-induced instability for low-frequency variability. J. Atmos. Sci., 67, 19471964, https://doi.org/10.1175/2009JAS3185.1.

  • Jin, F. F., L. Pan, and M. Watanabe, 2006a: Dynamics of synoptic eddy and low-frequency flow interaction. Part I: A linear closure. J. Atmos. Sci., 63, 16771694, https://doi.org/10.1175/JAS3715.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jin, F. F., L. Pan, and M. Watanabe, 2006b: Dynamics of synoptic eddy and low-frequency flow interaction. Part II: A theory for low-frequency modes. J. Atmos. Sci., 63, 16951708, https://doi.org/10.1175/JAS3716.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kanamitsu, M., W. Ebisuzaki, J. Woollen, S. K. Yang, J. Hnilo, M. Fiorino, and G. Potter, 2002: NCEP–DOE AMIP-II reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 16311643, https://doi.org/10.1175/BAMS-83-11-1631.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kim, H., S.W. Yeh, S. An, J. Park, B. Kim, and E. Baek, 2020: Arctic sea ice loss as a potential trigger for central Pacific El Niño events. Geophys. Res. Lett., 47, e2020GL087028, https://doi.org/10.1029/2020GL087028.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lai, A. W., M. Herzog, and H. F. Graf, 2015: Two key parameters for the El Niño continuum: Zonal wind anomalies and western Pacific subsurface potential temperature. Climate Dyn., 45, 34613480, https://doi.org/10.1007/s00382-015-2550-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Larson, S. M., Y. Okumura, K. Bellomo, and M. Breeden, 2022: Destructive interference of ENSO on North Pacific SST and North American precipitation associated with Aleutian low variability. J. Climate, 35, 35673585, https://doi.org/10.1175/JCLI-D-21-0560.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lau, N.-C., 1988: Variability of the observed midlatitude storm tracks in relation to low-frequency changes in the circulation pattern. J. Atmos. Sci., 45, 27182743, https://doi.org/10.1175/1520-0469(1988)045<2718:VOTOMS>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., E. Guilyardi, J. P. Boulanger, C. Menkes, P. Delecluse, P. Inness, J. Cole, and J. Slingo, 2004: Triggering of El Niño by westerly wind events in a coupled general circulation model. Climate Dyn., 23, 601620, https://doi.org/10.1007/s00382-004-0457-2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Li, C. Y., 1990: Interaction between anomalous winter monsoon in East Asia and El Niño events. Adv. Atmos. Sci., 7, 3646, https://doi.org/10.1007/BF02919166.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lian, T., D. Chen, Y. Tang, and Q. Wu, 2014: Effects of westerly wind bursts on El Niño: A new perspective. Geophys. Res. Lett., 41, 35223527, https://doi.org/10.1002/2014GL059989.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Liu, M., H. L. Ren, R. H. Zhang, S. Ineson, and R. Wang, 2021: ENSO phase-locking behavior in climate models: From CMIP5 to CMIP6. Environ. Res. Lett., 3, 031004, https://doi.org/10.1088/2515-7620/abf295.

    • Search Google Scholar
    • Export Citation
  • Luo, J. J., S. Masson, S. Behera, S. Shingu, and T. Yamagata, 2005: Seasonal climate predictability in a coupled OAGCM using a different approach for ensemble forecasts. J. Climate, 18, 44744497, https://doi.org/10.1175/JCLI3526.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Luo, J. J., R. Zhang, S. K. Behera, Y. Masumoto, F. Jin, R. Lukas, and T. Yamagata, 2010: Interaction between El Niño and extreme Indian Ocean dipole. J. Climate, 23, 726742, https://doi.org/10.1175/2009JCLI3104.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Luo, J. J., C. X. Yuan, W. Sasaki, S. K. Behera, Y. Masumoto, T. Yamagata, J.-Y. Lee, and S. Masson, 2016: Current status of intraseasonal-seasonal-to-interannual prediction of the Indo-Pacific climate. Indo-Pacific Climate Variability and Predictability, S. K. Behera and T. Yamagata, Eds., World Scientific, 63–107.

    • Crossref
    • Export Citation
  • Mantua, N .J., and S. R. Hare, 2002: The Pacific decadal oscillation. J. Oceanogr., 58, 3544, https://doi.org/10.1023/A:1015820616384.

  • McPhaden, M. J.,