1. Introduction
El Niño–Southern Oscillation (ENSO) is the most outstanding climate fluctuation at seasonal to interannual time scales. It has an enormous influence on the climate/weather not only locally but also remotely (McPhaden et al. 2006; Timmermann et al. 2018; Cai et al. 2021). Hence, various efforts have been made to understand the ENSO developing process through various methods, such as ENSO theory, climate model simulation, and observational analysis with statistical–dynamical method (Jin 1997; Li 1997; Chen et al. 2004; Tang et al. 2018). In this light, understanding of the influence of the interbasin interaction via atmospheric teleconnections has also been augmented (Cai et al. 2019; Wang 2019).
As for the interbasin interaction in association with ENSO development, diverse climate variability modes, such as the Atlantic Niño, Western Hemisphere warm pool, South Atlantic subtropical dipole, Indian Ocean dipole, and the north tropical Atlantic (NTA) mode, have been investigated (Rodríguez-Fonseca et al. 2009; Izumo et al. 2010; Ham et al. 2013; Park et al. 2018). Among them, the role of the NTA in ENSO development has received much attention in recent years (Ham et al. 2013), despite some discussion of its actual effects with respect to ENSO autocorrelation (as will be addressed in the last section) (Richter et al. 2021; Zhang et al. 2021).
Ham et al. (2013) first showed that the sea surface temperature (SST) anomaly (SSTA) warming (cooling) in the NTA in boreal spring induces (regulates) deep convection in the tropical Atlantic, which generates westward-propagating Rossby waves accompanying anomalous northeasterly (southeasterly) wind and precipitation reduction (enhancement) over the subtropical North Pacific. Due to the air–sea interactions [moist static energy (MSE) feedback, also known as moist enthalpy advection process] around the Pacific intertropical convergence zone (ITCZ) (Wu et al. 2017), these signals relay into the tropical Pacific, consequently triggering a La Niña (an El Niño) event in the subsequent winter. Simultaneously, the eastward Kelvin wave energy propagation along the equator plays an additional role in modulating the low-level zonal wind anomalies in association with the Walker circulation change in the equatorial western Pacific, contributing to ENSO development (Rong et al. 2010; Yu et al. 2016; Jiang and Li 2021).
A previous study demonstrated that the climate models participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5) tend to simulate the NTA effect on the following ENSO (Ham and Kug 2015). At the same time, CMIP5 models very diversely simulate the NTA effect on ENSO. Such diversity is partly caused by the different temporal evolution of ENSO (i.e., phase locking) in the CMIP5 models considering that the lagged NTA influence on ENSO reflects seasonality. For example, the climate model with the peak phase of ENSO in seasons other than winter tends to simulate the NTA effect on ENSO differently from that in observations. More importantly, climate models with wetter climatology in the equatorial Atlantic in boreal spring are argued to tend to have a strong impact of the NTA on ENSO. This argument is associated with the fact that the strong mean precipitation based on the abundant moisture provides better conditions for the generation of deep convection, and effective atmospheric teleconnections (Zebiak 1986; Watanabe et al. 2011; Ham and Kug 2012).
The abovementioned study attributes the NTA effect efficiency on ENSO in climate models to the local effect in the tropical Atlantic. Nevertheless, from the perspective of the NTA-induced signals able to grow by themselves via local ocean–atmosphere feedback once they reach the subtropical North Pacific, the importance of the role of the ocean–atmosphere feedback in the subtropical North Pacific cannot be ignored. For example, the wind–evaporation–SST (WES) feedback is enhanced during the period when the climatological mean trade wind intensifies, which efficiently conveys midlatitude signals to the tropics (Yu et al. 2012, 2015; Park et al. 2021). This result implies the importance of climatology over the subtropical North Pacific in the NTA–ENSO connection regarding the modulation of the ocean–atmosphere feedback.
In the meantime, observational studies have shown that the strength of the Atlantic effect on ENSO is not stationary but rather has interdecadal modulation (Rodríguez-Fonseca et al. 2009; Martín-Rey et al. 2014; Park and Li 2019). Particularly, the strength of the NTA effect on ENSO has been continuously intensifying since the 1960s (Wang et al. 2017). The warmer basic state in the NTA region after the 1990s likely due to the Atlantic multidecadal oscillation (AMO) phase change or global warming trends is more likely capable of channeling the Atlantic influence into the eastern subtropical Pacific. A warm Atlantic can also intensify the climatological subtropical North Pacific high, making the wind–evaporation–SST feedback strong. Consequently, the NTA effect easily transports over the subtropics (Yu et al. 2015).
As mentioned above, a difference exists in the NTA effect simulation on ENSO in CMIP5. In observation, the NTA effect on ENSO is nonstationary. In this context, we will first attempt to further examine what is responsible for the different NTA effect simulation on ENSO in CMIP5. We will then discuss whether or not what we found from the CMIP5 analysis can be applied to understand the nonstationary NTA effect in the observations. The remainder of this paper is organized as follows: section 2 introduces the CMIP5 and reanalysis datasets representing the observations utilized in this study, section 3 addresses the factors responsible for the diverse relationship between the NTA and ENSO in the CMIP5 climate models, section 4 presents observations on the nonstationary NTA effect on ENSO, and section 5 provides a summary and discussion.
2. Dataset and indices
The historical simulations of 30 climate models participating in the CMIP5 were analyzed in this study (Table 1; Taylor et al. 2012). The datasets were obtained online (https://esgf-node.llnl.gov/projects/cmip5). Only the first ensemble member was utilized. The analysis period was set to 1900–2000. SST, sea level pressure (SLP), low-level wind (at 850 hPa), and precipitation were used as the primary variables.
Thirty climate models used in this study, all of which participated in CMIP5. The model name, modeling center, and resolution are presented from left to right. All datasets were interpolated into 144 (longitude) × 73 (latitude) before further analysis. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)
The NTA and ENSO indices were first defined to investigate the strength of the NTA effect on ENSO. The ENSO index was represented herein by the Niño-3.4 index averaged from December to the following February (DJF) over the Niño-3.4 region (5°S–5°N, 120°–170°W). Areally averaged SSTA over 0°–15°N, 90°W–20°E in March–May (MAM) was conducted to obtain the NTA index. The previous winter ENSO signal was then linearly removed from the areal-averaged index based on a linear regression against the ENSO index to exclude the ENSO effect (refer to Ham et al. 2013). For both the NTA and ENSO indices, the slight season changes (e.g., February–April for NTA and November–January for ENSO) did not cause significant differences in the results. All datasets were interpolated such that the horizontal resolution was set to 2.5° × 2.5° before the analysis was conducted. The trends and climatological means were removed from all the variables.
For the observational reanalysis data, the monthly dataset of the Twentieth Century Reanalysis V3 (20CRv3) from the National Ocean Atmosphere Administration (NOAA), which is an assimilated dataset that uses a state-of-the-art analysis and forecast system, was utilized as the primary atmospheric dataset (Slivinski et al. 2019). This dataset was selected because of its long-term record spanning from 1836 to 2015. The SLP, wind, and precipitation data from this dataset were used. The main dataset for the SST is the Extended Reconstruction Sea Surface Temperature version 5 (ERSSTv5). It is a global monthly SST dataset derived from the International Comprehensive Ocean–Atmosphere Dataset (ICOADS) available from 1854 to the present. Both reanalysis datasets are available online (https://psl.noaa.gov/data/gridded/). To compare the biases among other datasets, the ERSSTv3 and the Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST) datasets were utilized (Rayner et al. 2003; Huang et al. 2017). The analyzed period in this study was from 1885 to 2015, covering the entire twentieth century. Hence, it seemed appropriate to examine the decadal-to-interdecadal modulation of the climate phenomena. The same processes applied to the CMIP5 climate model analysis were applied to obtain the NTA and ENSO indices in the observational analysis. Additionally, the AMO and interdecadal Pacific oscillation (IPO) indices are used here for comparison (Enfield et al. 2001; Henley et al. 2015), These indices were downloaded from NOAA (https://www.esrl.noaa.gov/psd/data/timeseries/AMO or https://www.esrl.noaa.gov/psd/data/timeseries/IPOTPI).
3. NTA effect on ENSO in the CMIP5 models
a. Strong and weak NTA–ENSO connection groups
The lagged correlation coefficient between the NTA and the ENSO indices with a three-season lag in each climate model was calculated to measure the strength of the lagged relationship between the NTA and ENSO in the CMIP5 models. Figure 1 illustrates the correlation coefficients in the 30 CMIP5 climate models. The climate models very diversely simulated the strength of the NTA effect on ENSO, which ranged from −0.42 (GFDL-ESM2M) to 0.28 (INMCM4) over 1900–2000. The coefficient for observations was −0.25 over 1885–2015. A positive correlation coefficient indicates that the NTA effect on ENSO simulated in the model is opposite to that in the observations. The majority of the climate models tended to underestimate the NTA effect on the ENSO when compared with those in the observational results.
Lagged correlation coefficients between the NTA index and the following ENSO index. The leftmost bar in dark gray indicates the observational result from 1885 to 2015. The second bar in brown indicates the multimodel ensemble mean from the 30 climate models participating in CMIP5 from 1900 to 2000. From the third-to-rightmost bar, the correlation coefficients from each climate model are shown in increasing order, wherein the model number is assigned according to the results. The dark blue and sky blue bars indicate the nine highest (top 30%) and nine lowest (bottom 30%) models representing the strong and weak NTA–ENSO connection models, respectively. The asterisk (*) indicates the 95% confidence level via the Student’s t test.
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
The nine highest (top 30%) and nine lowest (bottom 30%) climate models were selected and classified as the strong and weak NTA–ENSO connection groups, respectively, to understand the diversity of the NTA–ENSO connection among the CMIP5 climate models. The characteristics of the NTA effect on ENSO between the two groups were compared. Note that the present study results were not significantly sensitive to the number of both groups.
The anomalous SST, SLP, and low-level wind data were regressed onto the normalized NTA index in each climate model to compare the NTA effects on ENSO between the two groups. The results were then ensemble-averaged into the strong and the weak NTA–ENSO connection groups (Fig. 2). For the strong NTA–ENSO connection group (Fig. 2a), in spring [MAM(0)], the SSTA warming in the NTA was well observed, with a cyclonic circulation over the North Atlantic. The NTA SSTA warming penetrated to the equatorial far eastern Pacific accompanying enhanced precipitation along the equator from Atlantic to eastern Pacific. A cyclonic circulation over the subtropical eastern North Pacific synchronizes with the SSTA warming and the enhanced precipitation, which resulted from the Rossby wave propagations. The northerly wind anomalies over the western region of the cyclonic circulation conveyed a negative MSE to the tropics, causing a precipitation reduction over the Pacific ITCZ. This precipitation reduction generated off-equatorial Gill-type atmospheric responses (Gill 1980), which induced an anticyclonic circulation over the western region of the negative precipitation anomalies, namely over the western Pacific. This indicated the operation of the MSE feedback, also known as the moist enthalpy advection mechanism, over the subtropical North Pacific (Ham et al. 2013; Wu et al. 2017). Meanwhile, the enhanced precipitation along the equatorial Atlantic modulated the Walker circulation, contributing to generation of easterly wind in the equatorial western Pacific (Wang 2006; Ding et al. 2012; Frauen and Dommenget 2012; Kucharski et al. 2015; Polo et al. 2015; Martín-Rey et al. 2014).
(a) Ensemble of the strong NTA–ENSO connection models of the SSTA (shaded; shading bar at right; unit: °C), anomalous low-level wind (unit: m s−1; at 850 hPa), and anomalous precipitation (dots, green for positive and brown for negative) regressed onto the NTA index. Results are shown for (top to bottom) MAM(0), MJJ(0), JAS(0), and D(1)JF(2), wherein the SSTA, low-level wind, and precipitation are marked when 95% confidence level by the Student’s t test is satisfied. (b) As in (a), but for the weak NTA–ENSO connection models. (c) Differences between (a) and (b), wherein the SSTA, low-level wind, and precipitation are marked when 95% confidence level by bootstrap (1000 times) is satisfied.
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
In May–July [MJJ(0)] to July–September [JAS(0)], the anomalous positive precipitation over the equatorial Atlantic maintained its strength, despite the NTA–SSTA warming being somewhat decaying. The enhanced precipitation-induced Rossby wave propagation kept inducing the northerly wind anomalies and the precipitation reduction over the subtropical North Pacific. The easterly wind anomalies in the western equatorial Pacific further grew via both the modulated Walker circulation and the anticyclonic Rossby wave response to the negative precipitation over the subtropical North Pacific. La Niña then began to manifest through the Bjerknes feedback in the following winter.
As for the weak NTA–ENSO connection group (Fig. 2b), in MAM(0), the substantial SSTA warming in the NTA to the far equatorial eastern Pacific was observed with a cyclonic circulation in the North Atlantic. Nevertheless, the northerly wind anomalies over the eastern North Pacific and its associated SST cooling were not prominent, whereas the SST warming and westerly wind along the equatorial eastern Pacific were strong. These results were different from those in the case of the strong NTA–ENSO connection group (Fig. 2c). From MAM(0) to JAS(0), the precipitation in the equatorial Atlantic decreased as the NTA SSTA decayed. Over the equatorial eastern Pacific, the anomalous westerly wind seemed to maintain the SSTA warming in the eastern Pacific, which hindered La Niña development.
With respect to the difference between the two groups, the precipitation along the Atlantic ITCZ in the spring to summer season was stronger in the strong NTA–ENSO connection group than in the weak NTA–ENSO connection group, albeit the SSTA warming being comparable (Fig. 2c). In other words, both Rossby and Kelvin waves in association with the strong precipitation in the equatorial Atlantic played an important role in modulating the NTA–ENSO connection. This interpretation explains why the NTA warming can promote ENSO events in the strong NTA–ENSO connection models but not in the weak NTA–ENSO connection models.
b. Climatological mean state difference
The different response of the precipitation in the Atlantic to the NTA warming was obvious between the strong and weak NTA–ENSO connection groups (Fig. 2c). The standard deviation of the NTA index explains the approximately 13% (r = −0.36) intermodel variability of the NTA–ENSO connection, which is significant at a 90% confidence level. Apart from the difference in precipitation in the Atlantic, prominent differences were also observed in the subtropical North Pacific from spring to summer. That is, the northeasterly wind anomalies were widely observed over the subtropical North Pacific, with the negative precipitation anomalies along the Pacific ITCZ. These negative precipitation anomalies further developed with time. These results motivated us to speculate the role of the subtropical North Pacific mean climatology in the different strengths of the NTA–ENSO connection in climate models.
Based on this argument, we aimed to understand how the climatological mean states differ between the strong and weak NTA–ENSO connection groups. Accordingly, the climatological mean SST, low-level wind, SLP, and precipitation in April–June were obtained from each model. These were then regressed against the correlation coefficients between the NTA and ENSO indices in the intermodel space (Fig. 1). April–June was selected because the ocean–atmosphere coupling feedback could efficiently operate after the NTA effect in boreal spring is transported to the North Pacific in the climate models. Note that a slight season change does not make a significant difference in the results.
Figure 3a shows that significant signals can be mostly found in the subtropical North Pacific, where the signals were well organized compared with those in the tropical Atlantic. For the strong NTA–ENSO connection group, we observed a warmer SST with a westerly wind and large precipitation in the equatorial far eastern Pacific (0°–15°N, 90°–120°W), which seemed to play a role in linking the NTA signals to the Pacific (Ham and Kug 2015). The strong subtropical North Pacific high system is observed well, including the northeasterly trade wind, cool SST, and weak precipitation in its southeastern part. Herein, the strong regions of the SLP (15°–30°N, 150°W–180°), low-level zonal wind (5°–15°N, 150°E–150°W), and precipitation (12.5°–17.5°N, 170°E–140°W) are marked by the red boxes in Fig. 3a. Figures 3b and 3c show the close relationship between the climatological SLP and the zonal wind and between the climatological zonal wind and the precipitation over the subtropical North Pacific in April–June in association with the subtropical North Pacific high system. This result implied that the climatological subtropical North Pacific high somehow contributed to the determination of the strength of the NTA–ENSO connection in the climate model.
(a) Regression maps of the climatological mean [April–June (AMJ)] SST (shading), SLP (contours), low-level wind (vectors), and precipitation (dots, green for positive and brown for negative) of models from CMIP5 against the NTA–ENSO connection index in the intermodel space (Fig. 1). Low-level wind and precipitation are marked when significant at a 95% confidence level using the Student’s t test. Gray hatching indicates that the SLP signals are significant at the 95% confidence level using the Student’s t test. The subtropical high core regions for the SLP (15°–30°N, 150°W–180°), precipitation (12.5–17.5°N, 170°E–140°W), and trade winds (5°–15°N, 150°E–150°W) are depicted by the top, middle, and bottom red boxes, respectively. (b) Scatterplot between the climatological (AMJ) SLP (15°–30°N, 150°W–180°) and the climatological zonal wind (5°–15°N, 150°E–150°W) in the CMIP5 intermodel space, in which the slope of the dotted line shows the regression coefficient. The model numbering denotes the order of SFM efficiency given in Fig. 1. The strong and weak NTA–ENSO connection models are identified by the dark blue and sky blue colors, respectively. The R value indicates the correlation coefficient and the asterisks (**) denote a 99% confidence level using the Student’s t test. (c) As in (b), but between the climatological zonal wind and the climatological precipitation (12.5°–17.5°N, 170°E–140°W).
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
We further examined the relationship between the subtropical North Pacific climatology and the strength of the NTA–ENSO connection in the intermodel space (Fig. 4). Figure 4a shows the relationship between the climatological SLP in the AMJ and the strength of the NTA–ENSO connection in the intermodel space. The correlation coefficient between them is −0.51, which was significant at the 99% confidence level using the Student’s t test. We also examined the relationship between the climatological mean precipitation (zonal wind) and the strength of the NTA–ENSO connection, in which we confirmed a significant relationship at the 99% confidence level (correlation coefficient = 0.52 for both precipitation and zonal wind) (Figs. 4b,c). These results indicate that the NTA–ENSO connection tends to be proportional to the intensity of the subtropical North Pacific high in the spring season. Meanwhile, the correlation coefficient between the climatological mean precipitation in the NTA and the NTA–ENSO connection strength was 0.43, which was significant at the 95% confidence level (not shown). In summary, the local mean precipitation over the NTA can play a significant role in the NTA–ENSO connection; however, it is less than that with the subtropical North Pacific high, implying that climatology over the subtropical Pacific plays a more crucial role than that over the tropical Atlantic.
(a) Scatterplot between the climatological (AMJ) SLP (15°–30°N, 150°W–180°) and the NTA–ENSO correlation coefficients (Fig. 1) in the CMIP5 intermodel space. The dotted line slope shows the regression coefficient. The model numbering denotes the SFM efficiency order given in Fig. 1. The strong and weak NTA–ENSO connection models are identified by the dark blue and sky blue colors, respectively. The R value indicates the correlation coefficient and the asterisks (**) mean the 99% confidence level using the Student’s t test. (b) As in (a), but between climatological precipitation (12.5°–17.5°N, 170°E–140°W) and NTA–ENSO correlation coefficients. (c) As in (a), but between the climatological zonal wind (5°–15°N, 150°E–150°W) and NTA–ENSO correlation coefficients.
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
Next, we examined how the subtropical North Pacific high system modulated the NTA–ENSO connection. To answer this question, we must address the air–sea coupled feedbacks (i.e., MSE feedback). In terms of the MSE feedback, the NTA–SSTA warming formed the westward propagation of the Rossby waves, which induced the northerly wind anomalies over the subtropical North Pacific. The northerly wind anomalies transported a negative MSE into the Pacific ITCZ; hence, the precipitation along the Pacific ITCZ decreased. Its relevant Gill-type atmospheric responses then induced northeasterly wind anomalies over the western area of the negative precipitation anomalies, which again caused precipitation reduction along the ITCZ. During the spring to summer seasons, the Pacific ITCZ moves poleward. Such feedback processes efficiently operate over the subtropical North Pacific. In association with the MSE feedback, the northerly wind anomaly with the weak climatological mean precipitation along the subtropics can convey a more negative MSE into the Pacific ITCZ, consequently accelerating the precipitation reduction. Herein, the cold mean SST played some secondary role in the negative MSE formation (Wu et al. 2017).
Figure 5a illustrates the relationship between the climatological precipitation amount (12.5°–17.5°N, 170°E–140°W) and the precipitation response (5°–15°N, 160°E–130°W) to the unit change of the meridional wind (10°–20°N, 160°E–130°W) in the AMJ to represent the MSE feedback [d(PRa)/d(Va)] around the Pacific ITCZ. The meridional wind component was used because of its direction perpendicular to the longitudinal spread of the precipitation associated with the ITCZ (Fig. 3a). They were negatively correlated with each other at the 99% confidence level (correlation coefficient: −0.81), implying that a lesser amount of climatological precipitation on the poleward side of the Pacific ITCZ [enhanced meridional gradient in PR; −d(Pr)/dy)] favored the efficient MSE feedback. The intensity of the MSE feedback over the subtropical Pacific tends to explain the strength of the NTA–ENSO connection (Fig. 5b).
(a) Scatterplot between climatological precipitation (12.5–17.5°N, 170°E–140°W) and precipitation response (5°–15°N, 160°E–130°W) to the unit change in meridional wind (10°–20°N, 160°E–130°W) in the CMIP5 intermodel space. The dotted line slope shows the regression coefficient. The model numbering denotes the order of SFM efficiency given in Fig. 1. The strong and weak NTA–ENSO connection models are identified by the green and yellow colors, respectively. The R value indicates the correlation coefficient and asterisks (**) denote the 99% confidence level using the Student’s t test. (b) As in (a), but between the precipitation response to the unit change in the meridional wind and the NTA–ENSO correlation coefficient (Fig. 1).
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
The previous studies discussed that the WES feedback efficiently occurred in the subtropical North Pacific (Yu et al. 2015). We examined this by defining the WES feedback parameter as d(SSTa)/d(Ua) over 5°–15°N, 150°E–150°W in April–June (AMJ), wherein the zonal wind was focused on by considering the main trade wind direction. The SSTA in the subtropical Pacific increased with the weakened trades in the almost all climate models (29 models), indicating a positive WES feedback. The relationship between the WES feedback in the subtropical Pacific and the NTA–ENSO connection strength was weak (Fig. 5c). This result implies that the WES feedback seems irrelevant to the NTA–ENSO connection from an intermodel perspective.
4. NTA effect on ENSO in observation
a. NTA effect on ENSO and its interdecadal modulation
We looked into the NTA effect on ENSO in observations. Figure 6 shows the regressed SSTA, SLP anomaly (SLPA), anomalous low-level wind, and anomalous precipitation against the NTA index from spring to the following winter over the last century. In spring [MAM(0)], we observed a strong NTA–SSTA warming with anomalous cyclonic circulation in the North Atlantic. Concurrently, precipitation occurred along the Atlantic to the far eastern Pacific, consequently inducing Rossby wave propagation toward the subtropical eastern North Pacific. The wave propagation led to anomalous northerly wind and precipitation reduction along the Pacific ITCZ in spring to summer. Through air–sea coupled feedback processes, easterly wind anomalies developed along the equatorial Pacific, leading to La Niña in the subsequent winter.
Regressed SSTA (shading; °C; color bar at right), low-level wind anomaly (vectors; at 850 hPa), and precipitation anomaly (dots; green for positive and brown for negative) against the NTA index during 1885–2015, for (from top to bottom) the March–May (MAM), May–July (MJJ), July–September (JAS), and December–February (DJF) means, respectively. The SSTA, winds, and precipitation are drawn when the 95% confidence level using the Student’s t test is satisfied.
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
According to the previous observational study (Wang et al. 2017), the NTA effect on ENSO is not stationary. The correlation analysis between NTA and ENSO indices with a 30-yr moving window was conducted for the period of 1885–2015 to further examine the nonstationary state of the NTA effect (Fig. 7a). In this figure, the NTA effect on ENSO gets intensified from 1960 to 2015, which is consistent with the previous results. Additionally, the NTA effect on ENSO was strong around the 1920s. The result verifies that the NTA effect on ENSO has a prominent interdecadal modulation. We also utilized ERSSTv3 and HadISST reanalysis and obtained consistent results.
(a) 30-yr moving correlation coefficients between the NTA and ENSO indices with a three-season lag (left axis; black) in ERSSTv5 (red), ERSSTv3 (black, short-dashed), and HadISST (black, long-and-short dashed). The dot–dashed line indicates the guideline of 95% confidence level, −0.36. (b) 30-yr February–May (FMAM) moving averages of the subtropical zonal wind (10°–20°N, 150°W–180°; blue) and subtropical precipitation (15°–30°N, 150°W–180°; green). (c) Smoothed AMO (orange) and IPO (aqua) index from NOAA.
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
We examined the interdecadal modulation of the climatological mean states to understand the interdecadal modulation of the NTA effect. To do this, the interdecadal modulation of the climatological mean state was obtained by the moving average with a 30-yr window regressed against the 30-yr moving averaged NTA–ENSO index (Fig. 7a). As a climatological mean state, we focused on February–May because the subtropical North Pacific high is prominent during this season. Figure 8 illustrates the regression results, in which a subtropical high system was well observed in the subtropical North Pacific. In the observations, the NTA effect on ENSO gets intensified during the decades when the subtropical North Pacific high system was well developed. This interpretation is consistent with the results obtained from the CMIP5 climate models analysis, in which the climate models with a strong subtropical high tend to have a strong NTA effect on ENSO, despite the region of the subtropical high system being limited to a small region (15°–30°N, 150°E–180°; Fig. 8), compared to the CMIP5 result (Fig. 3a).
Regression maps of the long-term (30-yr) climatological mean (February–May) SST (shading), SLP (contours), low-level wind (vectors), and precipitation (dots; green dots for positive and brown for negative) with a 30-yr moving NTA–ENSO connection index (Fig. 7a). The low-level wind and precipitation are marked only when significant at the 95% confidence level using the Student’s t test.
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
Strong trade winds and less precipitation were observed around the southeastern area of the subtropical North Pacific high; that is, the NTA effect depends on the trade wind intensity. Based on these findings, we examined the interdecadal modulation of the trade wind intensity (10°–20°N, 150°W–180°) and precipitation (15°–30°N, 150°W–180°) by its moving average with a 30-yr window (Fig. 7b). The interdecadal modulation of the trade wind intensity and precipitation matched well with the NTA effect on ENSO from the interdecadal time scales, despite some discrepancy. This result implies the importance of the subtropical North Pacific high in the NTA effect efficiency on the tropical Pacific.
b. Long-term natural variability and interdecadal modulation of the NTA effect on ENSO
In the observations, the interdecadal variation of the subtropical North Pacific high can largely explain that of the NTA effect on ENSO (Fig. 8). Hence, discussing whether or not the interdecadal modulation of the subtropical North Pacific high can be modulated by the well-known natural variability with a low frequency (e.g., AMO and IPO) would be beneficial. Nevertheless, both the AMO and the IPO seem to have different movements with the NTA effect on ENSO from the interdecadal time scales (Fig. 7c). During the recent decades (1960–2000), the AMO and the IPO particularly showed an oscillatory behavior, whereas the NTA effect on ENSO continuously intensified. The results seem to imply the existence of an interdecadal signal that cannot be fully explained by the AMO/IPO. However, the role of the AMO/IPO in the NTA–ENSO connection is complicated with diverse spatiotemporal scales; thus, it does not necessarily mean that the AMO/IPO plays an insignificant role in the modulation of the NTA–ENSO connection.
As for the other natural long-term variabilities associated with the trade wind over the subtropical North Pacific, note the Aleutian low located over the northern area of the subtropical high (Fig. 8). The subtropical North Pacific high coexists with the Aleutian low as a counterpart; in other words, they can be considered as a unified system. An empirical orthogonal function (EOF) analysis was applied to the SLPA over the North Pacific to investigate the subtropical North Pacific high system in association with the Aleutian low. The first EOF mode showed the Aleutian low variability mode, wherein the strong low SLPA is located over the southern area of the Aleutian Islands. The high SLPA is located in the southern area of the low SLPA, although its signal seems weak because the SLP variability weakens as it goes to the tropics from the midlatitudes.
To investigate the interdecadal modulation of the subtropical high with the Aleutian low, its principal component time series was 30-yr-moving-averaged (Fig. 9). Overall, the interdecadal modulation of the first mode intensity tended to match well with the NTA effect on ENSO until the 1990s (correlation coefficient: 0.81). This result implied the possibility of the NTA effect being dependent on the natural variability of the subtropical high associated with the Aleutian low on interdecadal time scales by modulating the trade wind intensity. However, after the 1990s, the NTA effect on ENSO kept intensifying, whereas the Aleutian low became weak. Shortly after 1990, the trade wind was not significantly related to the Aleutian low, indicating that the natural variability of the Aleutian low system cannot explain the NTA effect on ENSO any longer. We infer that the recent warming of the Atlantic can intensify the trade wind, allowing the NTA effect to become strong (i.e., global warming hiatus; McGregor et al. 2014; England et al. 2014; Meehl et al. 2021; Yang et al. 2020). In short, the NTA effect could be modulated by both the natural variability mode (subtropical high and Aleutian low) and external forcing (e.g., global warming). Hence, a relevant study is worth pursuing in the future.
(top) First EOF mode of the SLPA over the North Pacific during February to April. (bottom) The blue line indicates the 30-yr moving average of its principal component time series, whereas the black line depicts the same as the 30-yr moving correlation between the NTA and ENSO indices (Fig. 7a).
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
5. Summary and discussion
In this study, we revisited the diverse NTA effect on ENSO among the CMIP5 climate models and aimed to understand what is responsible for it, in view of remote response of the Pacific, rather than tropical Atlantic local forcing (Wang et al. 2017). We found that climate models with strong trade wind and less precipitation associated with the subtropical North Pacific high system tend to have an intensified NTA effect on ENSO because the MSE feedbacks, also known as the moist enthalpy advection mechanism, can be efficiently operated under the increased negative meridional gradient of moisture around the northern area of the Pacific ITCZ. Consequently, the NTA-induced signals can be developed well and conveyed to the deep tropical Pacific. In the observations, consistent with the CMIP5 analysis results, the NTA effect on ENSO was confirmed to intensify during the decades when the trade wind was strong, and subtropical precipitation was decreased. We believe that this result has a novelty in presenting the importance of the air–sea coupling process in the subtropical North Pacific apart from the Atlantic local effect (Ham and Kug 2015).
A discussion regarding the actual NTA influence on ENSO is ongoing (Richter et al. 2021; Zhang et al. 2021). This discussion is mainly attributed to two facts: 1) El Niño can strongly induce SSTA warming in the NTA in the following spring and 2) El Niño tends to be followed by La Niña (Enfield and Mayer 1997; Klein et al. 1999; Chiang and Sobel 2002; Chiang and Lintner 2005; Park and Li 2019; Chen et al. 2021). These facts complicate the actual influence of the NTA on the following ENSO. In this light, Zhang et al. (2021) argued that the NTA effect on ENSO is spurious due to the reflection of ENSO autocorrelation through the CMIP6 analysis. By analyzing the atmospheric general circulation model, Richter et al. (2021) also showed that the tropical Atlantic effect on ENSO was small. However, they also noted that the NTA effect on ENSO was significant, regardless of the previous ENSO, because their experiment eliminated ENSO variability, which was not subjected to ENSO autocorrelation. They also mentioned that the Pacific wind response to the NTA, which helped trigger ENSO, might be amplified by the coupled ocean–atmosphere process. Meanwhile, Ma et al. (2020) showed the possibility that the NTA SST anomaly acts as a precursor for ENSO by analyzing the ensemble hindcasts of the European Centre for Medium-Range Weather Forecasts (ECMWF) coupled model. Jiang and Li (2021) and Ma et al. (2022) also suggested that the NTA-related SSTA pattern must be examined to understand the exact effect of the NTA on the ENSO. These results implied that more studies are needed to better understand the NTA effect on the ENSO.
We also conducted the SSTA composite analysis for the NTA events with and without the previous ENSO to further discuss the role of the previous ENSO in the NTA–ENSO connection. In Fig. 10, notice that the NTA warming with the previous winter ENSO is greater than that without the previous winter ENSO due to the strong ENSO effect on the NTA. Nevertheless, the NTA effect on the equatorial Pacific was stronger without the previous ENSO. The composite result implies that the NTA can trigger ENSO events or at least help ENSO to further develop, regardless of its autocorrelation.
(a) Time-sequential SSTA composites of the NTA events with the previous ENSO events: 15 cases [i.e., 1886, 1888, 1889, 1915, 1958, 1969, 2005, and 2010 for warm NTA (>1σ) with the previous El Niño (>0.5σ) and 1904, 1911, 1923, 1974, 1975, 1976, and 1985 for the cold NTA (<−1σ) with the previous La Niña (<−0.5σ)]. (b) SSTA composite of the NTA events without the previous ENSO events: 13 cases [i.e., 1891, 1936, 1937, 1944, and 1963 for the warm NTA (>1σ) without the previous ENSO and 1908, 1913, 1922, 1930, 1947, 1986, 1991, and 1994 for the cold NTA (<−1σ) without the previous ENSO]. Composite results are displayed based on the warm NTA events (i.e., warm NTA minus cold NTA).
Citation: Journal of Climate 35, 20; 10.1175/JCLI-D-21-0933.1
Our study showed that the observational NTA effect on ENSO depends on the interdecadal modulation of the long-term climatology, and it is diverse among the CMIP5 climate models according to their climatology. Here, it is worthwhile to briefly mention the results from the climate models participating in CMIP6. We found that tropical teleconnection, rather than teleconnection via the subtropical North Pacific, seems much more important in the NTA–ENSO connection in CMIP6. We believe that the enhanced tropical teleconnection in CMIP6 models is closely related to the climatological significant SST warming in the western tropical Pacific. Comprehensively, a better simulation of the climatology in the climate models seems important in simulating the exact effect of the NTA on ENSO. As long as the majority of climate models suffer from simulating the correct climatology over the tropics, it would be difficult to exactly understand the NTA–ENSO interaction. Hence, more studies in association with climatology and its effect on the NTA–ENSO interaction should be pursued in the future.
Acknowledgments.
We appreciate the constructive comments and suggestions provided by the anonymous reviewers and the editor (Prof. Yuko M Okumura). This work was supported by the National Research Foundation of Korea (NRF) grant funded by the South Korean government (MSIT) (NRF-2020R1C1C1006569). This work was partly supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (NRF-2022R1A3B1077622). The authors declare no competing interests.
Data availability statement.
The SST dataset (Extended Reconstructed SST v3 and v5) and the atmospheric dataset (Twentieth Century Reanalysis v3) were downloaded from the National Centers for Environmental Prediction (https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/netcdf/). HadISST data were downloaded from https://www.metoffice.gov.uk/hadobs/hadisst/data/download.html. In this study, historical simulation datasets that participated in phase 5 of the Coupled Model Intercomparison Project (CMIP5) were downloaded from https://esgf-node.llnl.gov/projects/cmip5.
REFERENCES
Cai, W., and Coauthors, 2019: Pantropical climate interactions. Science, 363, eaav4236, https://doi.org/10.1126/science.aav4236.
Cai, W., and Coauthors, 2021: Changing El Niño–Southern Oscillation in a warming climate. Nat. Rev. Earth Environ., 2, 628–644, https://doi.org/10.1038/s43017-021-00199-z.
Chen, D., M. A. Cane, A. Kaplan, S. E. Zebiak, and D. Huang, 2004: Predictability of El Niño over the past 148 years. Nature, 428, 733–736, https://doi.org/10.1038/nature02439.
Chen, H.-C., F.-F. Jin, and L. Jiang, 2021: The phase-locking of tropical North Atlantic and the contribution of ENSO. Geophys. Res. Lett., 48, e2021GL095610, https://doi.org/10.1029/2021GL095610.
Chiang, J. C. H., and A. H. Sobel, 2002: Tropical tropospheric temperature variations caused by ENSO and their influence on the remote tropical climate. J. Climate, 15, 2616–2631, https://doi.org/10.1175/1520-0442(2002)015<2616:TTTVCB>2.0.CO;2.
Chiang, J. C. H., and B. R. Lintner, 2005: Mechanisms of remote tropical surface warming during El Niño. J. Climate, 18, 4130–4149, https://doi.org/10.1175/JCLI3529.1.
Ding, H., N. S. Keenlyside, and M. Latif, 2012: Impact of the equatorial Atlantic on the El Niño Southern Oscillation. Climate Dyn., 38, 1965–1972, https://doi.org/10.1007/s00382-011-1097-y.
Enfield, D. B., and D. A. Mayer, 1997: Tropical Atlantic sea surface temperature variability and its relation to El Niño–Southern Oscillation. J. Geophys. Res., 102, 929–945, https://doi.org/10.1029/96JC03296.
Enfield, D. B., A. M. Mestas-Nuñez, and P. J. Trimble, 2001: The Atlantic multidecadal oscillation and its relation to rainfall and river flows in the continental U.S. Geophys. Res. Lett., 28, 2077–2080, https://doi.org/10.1029/2000GL012745.
England, M. H., and Coauthors, 2014: Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nat. Climate Change, 4, 222–227, https://doi.org/10.1038/nclimate2106.
Frauen, C., and D. Dommenget, 2012: Influences of the tropical Indian and Atlantic Oceans on the predictability of ENSO. Geophys. Res. Lett., 39, L02706, https://doi.org/10.1029/2011GL050520.
Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106, 447–462, https://doi.org/10.1002/qj.49710644905.
Ham, Y.-G., and J.-S. Kug, 2012: How well do current climate models simulate two types of El Niño? Climate Dyn., 39, 383–398, https://doi.org/10.1007/s00382-011-1157-3.
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, 3103–3117, https://doi.org/10.1007/s00382-015-2527-z.
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, 112–116, https://doi.org/10.1038/ngeo1686.
Henley, B. J., J. Gergis, D. J. Karoly, S. Power, J. Kennedy, and C. K. Folland, 2015: A tripole index for the interdecadal Pacific oscillation. Climate Dyn., 45, 3077–3090, https://doi.org/10.1007/s00382-015-2525-1.
Huang, B., and Coauthors, 2017: Extended Reconstructed Sea Surface Temperature, version 5 (ERSSTv5): Upgrades, validations, and intercomparisons. J. Climate, 30, 8179–8205, https://doi.org/10.1175/JCLI-D-16-0836.1.
Izumo, T., and Coauthors, 2010: Influence of the state of the Indian Ocean Dipole on the following year’s El Niño. Nat. Geosci., 3, 168–172, https://doi.org/10.1038/ngeo760.
Jiang, L., and T. Li, 2021: Impacts of tropical North Atlantic and equatorial Atlantic SST anomalies on ENSO. J. Climate, 34, 5635–5655, https://doi.org/10.1175/JCLI-D-20-0835.1.
Jin, F.-F., 1997: An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual model. J. Atmos. Sci., 54, 811–829, https://doi.org/10.1175/1520-0469(1997)054<0811:AEORPF>2.0.CO;2.
Klein, S. A., B. J. Soden, and N.-C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932, https://doi.org/10.1175/1520-0442(1999)012<0917:RSSTVD>2.0.CO;2.
Kucharski, F., F. S. Syed, A. Burhan, I. Farah, and A. Gohar, 2015: Tropical Atlantic influence on Pacific variability and mean state in the twentieth century in observations and CMIP5. Climate Dyn., 44, 881–896, https://doi.org/10.1007/s00382-014-2228-z.
Li, T., 1997: Phase transition of the El Niño–Southern Oscillation: A stationary SST mode. J. Atmos. Sci., 54, 2872–2887, https://doi.org/10.1175/1520-0469(1997)054<2872:PTOTEN>2.0.CO;2.
Ma, J., W. He, Z. Chen, Y. Fu, and J. Yin, 2020: The impact of north tropical Atlantic sea surface temperature anomalies in the ensuing spring of El Niño on the tropical Indian Ocean and northwest Pacific. Int. J. Climatol., 40, 4978–4991, https://doi.org/10.1002/joc.6500.
Ma, J., H. Xu, J.-J. Luo, and S. Chen, 2022: Impact of tropical Atlantic SST anomaly on ENSO in the NUIST-CFS1.0 hindcasts. Int. J. Climatol., https://doi.org/10.1002/joc.7577, in press.
Martín-Rey, M., B. Rodríguez-Fonseca, I. Polo, and F. Kucharski, 2014: On the Atlantic–Pacific Niños connection: A multidecadal modulated mode. Climate Dyn., 43, 3163–3178, https://doi.org/10.1007/s00382-014-2305-3.
McGregor, S., A. Timmermann, M. F. Stuecker, M. H. England, M. Merrifield, F.-F. Jin, and Y. Chikamoto, 2014: Recent Walker circulation strengthening and Pacific cooling amplified by Atlantic warming. Nat. Climate Change, 4, 888–892, https://doi.org/10.1038/nclimate2330.
McPhaden, M. J., S. E. Zebiak, and M. H. Glantz, 2006: ENSO as an integrating concept in Earth science. Science, 314, 1740–1745, https://doi.org/10.1126/science.1132588.
Meehl, G. A., and Coauthors, 2021: Atlantic and Pacific tropics connected by mutually interactive decadal-timescale processes. Nat. Geosci., 14, 36–42, https://doi.org/10.1038/s41561-020-00669-x.
Park, J.-H., and T. Li, 2019: Interdecadal modulation of El Niño–tropical North Atlantic teleconnection by the Atlantic multi-decadal oscillation. Climate Dyn., 52, 5345–5360, https://doi.org/10.1007/s00382-018-4452-4.
Park, J.-H., J.-S. Kug, T. Li, and S. K. Behera, 2018: Predicting El Niño beyond 1-year lead: Effect of the Western Hemisphere warm pool. Sci. Rep., 8, 14957, https://doi.org/10.1038/s41598-018-33191-7.
Park, J.-H., M.-K. Sung, Y.-M. Yang, J. Zhao, S.-I. An, and J.-S. Kug, 2021: Role of the climatological intertropical convergence zone in the seasonal footprinting mechanism of El Niño–Southern Oscillation. J. Climate, 34, 5243–5256, https://doi.org/10.1175/JCLI-D-20-0809.1.
Polo, I., M. Martin-Rey, B. Rodriguez-Fonseca, F. Kucharski, and C. R. Mechoso, 2015: Processes in the Pacific La Niña onset triggered by the Atlantic Niño. Climate Dyn., 44, 115–131, https://doi.org/10.1007/s00382-014-2354-7.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.
Richter, I., H. Tokinaga, Y. Kosaka, T. Doi, and T. Kataoka, 2021: Revisiting the tropical Atlantic influence on El Niño–Southern Oscillation. J. Climate, 34, 8533–8548, https://doi.org/10.1175/JCLI-D-21-0088.1.
Rodríguez-Fonseca, B., I. Polo, J. García-Serrano, T. Losada, E. Mohino, C. R. Mechoso, and F. Kucharski, 2009: Are Atlantic Niños enhancing Pacific ENSO events in recent decades? Geophys. Res. Lett., 36, L20705, https://doi.org/10.1029/2009GL040048.
Rong, X., R. Zhang, and T. Li, 2010: Impacts of Atlantic sea surface temperature anomalies on Indo-East Asian summer monsoon–ENSO relationship. Chin. Sci. Bull., 55, 2458–2468, https://doi.org/10.1007/s11434-010-3098-3.
Slivinski, L. C., and Coauthors, 2019: Towards a more reliable historical reanalysis: Improvements for version 3 of the Twentieth Century Reanalysis system. Quart. J. Roy. Meteor. Soc., 145, 2876–2908, https://doi.org/10.1002/qj.3598.
Tang, Y., and Coauthors, 2018: Progress in ENSO prediction and predictability study. Natl. Sci. Rev., 5, 826–839, https://doi.org/10.1093/nsr/nwy105.
Taylor, K. E., R. J. Stouffer, and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485–498, https://doi.org/10.1175/BAMS-D-11-00094.1.
Timmermann, A., and Coauthors, 2018: El Niño–Southern Oscillation complexity. Nature, 559, 535–545, https://doi.org/10.1038/s41586-018-0252-6.
Wang, C., 2006: An overlooked feature of tropical climate: Inter-Pacific-Atlantic variability. Geophys. Res. Lett., 33, L12702, https://doi.org/10.1029/2006GL026324.
Wang, C., 2019: Three-ocean interactions and climate variability: A review and perspective. Climate Dyn., 53, 5119–5136, https://doi.org/10.1007/s00382-019-04930-x.
Wang, L., J.-Y. Yu, and H. Paek, 2017: Enhanced biennial variability in the Pacific due to Atlantic capacitor effect. Nat. Commun., 8, 14887, https://doi.org/10.1038/ncomms14887.
Watanabe, M., M. Chikira, Y. Imada, and M. Kimoto, 2011: Convective control of ENSO simulated in MIROC. J. Climate, 24, 543–562, https://doi.org/10.1175/2010JCLI3878.1.
Wu, B., T. Zhou, and T. Li, 2017: Atmospheric dynamic and thermodynamic processes driving the western North Pacific anomalous anticyclone during El Niño. Part II: Formation processes. J. Climate, 30, 9637–9650, https://doi.org/10.1175/JCLI-D-16-0495.1.
Yang, Y. M., S. Il An, B. Wang, and J. H. Park, 2020: A global-scale multidecadal variability driven by Atlantic multidecadal oscillation. Natl. Sci. Rev., 7, 1190–1197, https://doi.org/10.1093/nsr/nwz216.
Yu, J. H., T. Li, Z. M. Tan, and Z. W. Zhu, 2016: Effects of tropical North Atlantic SST on tropical cyclone genesis in the western North Pacific. Climate Dyn., 46, 865–877, https://doi.org/10.1007/s00382-015-2618-x.
Yu, J.-Y., M.-M. Lu, and S. T. Kim, 2012: A change in the relationship between tropical central Pacific SST variability and the extratropical atmosphere around 1990. Environ. Res. Lett., 7, 034025, https://doi.org/10.1088/1748-9326/7/3/034025.
Yu, J.-Y., P.-K. Kao, H. Paek, H.-H. Hsu, C.-W. Hung, M.-M. Lu, and S.-I. An, 2015: Linking emergence of the central Pacific El Niño to the Atlantic multidecadal oscillation. J. Climate, 28, 651–662, https://doi.org/10.1175/JCLI-D-14-00347.1.
Zebiak, S. E., 1986: Atmospheric convergence feedback in a simple model for El Niño. Mon. Wea. Rev., 114, 1263–1271, https://doi.org/10.1175/1520-0493(1986)114<1263:ACFIAS>2.0.CO;2.
Zhang, W., F. Jiang, M. F. Stuecker, F.-F. Jin, and A. Timmermann, 2021: Spurious north tropical Atlantic precursors to El Niño. Nat. Commun., 12, 3096, https://doi.org/10.1038/s41467-021-23411-6.