1. Introduction
Characterized by a north–south sea level pressure (SLP) dipole, the North Pacific Oscillation (NPO) is recognized as a primary atmospheric mode over the North Pacific (Walker and Bliss 1932; Rogers 1981; Wallace and Gutzler 1981; Linkin and Nigam 2008). It has been reported that various climate/weather phenomena around the Pacific rim are involved in the NPO. For example, low-level winds accompanied by NPO are responsible for the formation of a meridional sea surface temperature (SST) anomaly (SSTA) tripole, known as the Victoria mode or North Pacific Gyre Oscillation, by modulating latent heat fluxes at the ocean surface (Bond et al. 2003; Di Lorenzo et al. 2008; Ceballos et al. 2009). The NPO also has an enormous influence on both the East Asian winter monsoon (Wang et al. 2007; Linkin and Nigam 2008) and extreme winter weather in North America, which have been frequently observed in the 2010s (Screen and Simmonds 2014; Baxter and Nigam 2015; Hartmann 2015; Sung et al. 2019). In addition, it has been suggested that the NPO modulates spring precipitation in Australia in the Southern Hemisphere through ocean–atmosphere interactions in the tropical Pacific (Song et al. 2016).
Another well-known influence of the NPO is the triggering of El Niño–Southern Oscillation (ENSO) with a 1-yr lag, the process of which is referred to as the seasonal footprinting mechanism (SFM; Vimont et al. 2001, 2003a,b; Alexander et al. 2010). The SFM manifests through trade winds modified by the NPO southern lobe during winter. When the trade winds are intensified (weakened), the ocean surface becomes cool (warm) via wind–evaporation–SST (WES) feedback (Xie and Philander 1994). WES feedback can become more pronounced from winter to summer as the intertropical convergence zone (ITCZ) migrates poleward (Amaya 2019; Amaya et al. 2019), relaying subtropical signals to the tropics and forming zonal wind anomalies from the western to central equatorial Pacific Ocean. The relevant zonal wind stress anomalies initiate equatorial oceanic Kelvin waves, which contribute to initiating ENSO development. Meanwhile, the NPO-relevant wind stress curl in the tropics modulates the subsurface heat content along the equator through Sverdrup transport, providing a favorable precondition for ENSO development (Jin 1997; Li 1997; Anderson et al. 2013; Anderson and Perez 2015).
However, it is known that not every NPO is accompanied by an ENSO event the following winter, implying complicated connections between the NPO and ENSO (Amaya 2019; Chen and Yu 2020; Zhao et al. 2020). Park et al. (2013) showed that about 40% of NPO events in their positive phase preceded an El Niño (for convenience, in this study, a positive-phase NPO corresponds to a north anticyclonic–south cyclonic dipolar SLP structure). They argued that the intensity of the surface net heat flux in the North Pacific determines whether or not the SFM will operate efficiently through ocean–atmosphere interactions.
Different approaches have been used to understand the efficiency of the SFM process. Some approaches have focused on the characteristics of NPOs, whereas others have considered interactions with other climate phenomena or background conditions. As for the former, it has been suggested that the meridional/zonal position of the NPO is critical in determining how the SFM operates (Chen and Wu 2018; Yeh et al. 2018). In addition, NPO intensity has been suggested to be one of the key factors for understanding SFM diversity when analyzing large ensemble simulations of the Canadian Earth System Model (CanESM2; Chen and Yu 2020). That is, the stronger the NPO is, the more efficiently the NPO triggers an ENSO event.
As for the latter, Chen et al. (2013, 2014) argued that the interannual springtime Arctic Oscillation (AO) also played a role in the dynamics between the NPO and ENSO. They showed that the positive phase of the NPO is readily followed by an El Niño event when the springtime AO is positive, as the AO modulates the interaction between synoptic-scale eddies, the mean flow, and the resulting transportation of vorticity in the North Pacific. Additionally, it has been shown that the Atlantic multidecadal oscillation (AMO) influences the winter NPO–ENSO relationship by modulating climatological precipitation over the tropical North Pacific (Chen et al. 2019).
These previous studies have attributed the observed diversity in SFM behaviors to NPO characteristics (i.e., intensity and pattern), interactions with other phenomena (e.g., AO), or background conditions (e.g., AMO), especially outside of the tropical Pacific. However, given the importance of the ocean–atmosphere interaction over the tropical Pacific for the SFM, it is necessary to examine tropical conditions as well. It should be noted that WES feedback, a key process in the SFM, is substantially affected by seasonal conditions in the tropical Pacific (e.g., background trade winds or ITCZ structure), as it determines (i) the sensitivity of evaporation at the ocean surface to low-level wind changes (Vimont et al. 2009) and (ii) the strength and structure of atmospheric responses to anomalous SST (Amaya 2019). In terms of i, it is known that the efficiency of WES feedback is proportional to the background trade wind strength, as stronger (weaker) background trade winds cause a greater downward (upward) anomalous latent heat flux for the same anomalous westerly (easterly) wind speed (Vimont et al. 2009; Vimont 2010). Therefore, in this study, it was important to examine the Pacific ITCZ position to understand the relationship between the NPO and subsequent ENSO events with respect to ii. In particular, we focused on how the latitudinal displacement of the climatological ITCZ, in association with the distribution of climatological precipitation around the ITCZ, modulated atmospheric responses to SSTA in the subtropics and eventually the SFM.
The NPO–ENSO relationship is known to vary considerably by decades in observational studies (Yeh et al. 2018). As will be shown in this study, climate models also exhibit great diversity when simulating NPO–ENSO relationships. To understand what controls the NPO–ENSO relationship, we investigated the role of the tropical mean state in the SFM process by employing models from phase 5 of the Coupled Model Intercomparison Project (CMIP5) and an observational dataset. The remainder of this paper is organized as follows. The climate model and the reanalysis of the observational dataset utilized in this study are introduced in section 2. SFM efficiency and its regulating factors in climate models and observational dataset are presented in section 3 and section 4, respectively. A summary and discussion are given in the last section.
2. Dataset and indices
In this study, historical runs (1900–2000, first ensemble member) of 30 climate models participating in the CMIP5 were analyzed (Table 1), and the datasets were obtained online (https://esgf-node.llnl.gov/projects/cmip5). To investigate the SFM, the ENSO and NPO indices were first defined. ENSO is represented by the Niño-3.4 index, averaged from November to the following February (NDJF). To obtain the NPO index, an empirical orthogonal function (EOF) analysis was applied to monthly SLP anomalies for the North Pacific domain (120°E−100°W, 20°–70°N), and then the seasonal (NDJF) average of the second principal component time series was applied. To exclude simultaneous ENSO effects, the ENSO signal was linearly removed from the NPO index based on a linear regression against the Niño-3.4 index. Without these steps, prominent ENSO-like signals in the tropics co-occur with the NPO in some models, causing a causality problem between the NPO and ENSO, as discussed by Yeh et al. (2018). Trends and climatological means were removed before conducting the analysis. For both the ENSO and NPO indices, slight changes in the season (e.g., from December to February or from November to January) did not cause significant differences in the results. All of the datasets were interpolated so that the horizontal resolution was set to 2.5° × 2.5° before conducting the analysis.
30 Climate models used in this study, all of which participated in CMIP5. Before further analysis, all datasets are interpolated into 144 (longitude) × 73 (latitude) resolution according to the NCEP-R1. Expansions of most model names and centers can be found online (https://www.ametsoc.org/PubsAcronymList)
In this study, NPO intensity with regard to SFM was estimated by the following two methods. First, NPO intensity was defined as the standard deviation of the NPO index. Second, since the southern lobe of the NPO is more involved in the SFM, NPO intensity was obtained from a regional average (150°E−150°W, 10°–45°N) of the NPO pattern (second EOF pattern of the winter SLP). The obtained NPO intensities were used to examine the relationship to SFM efficiency (section 3).
For the observational reanalysis data, the monthly dataset of the National Centers for Environmental Prediction Reanalysis 1 (NCEP-R1), which is an assimilated dataset that uses a state-of-the-art analysis and forecast system, was utilized as the primary atmospheric dataset (Kistler et al. 2001). This dataset has been widely used due to its reliability as well as its relatively long record spanning from 1948 to the present. From this dataset, SLP, wind, and precipitation data were used. For SST, the Extended Reconstruction Sea Surface Temperature version 5 (ERSSTv5) was used (Huang et al. 2017). It is a global monthly SST dataset derived from the International Comprehensive Ocean–Atmosphere Data Set (ICOADS), available from 1854 to the present. Both reanalysis datasets are available online (https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/netcdf/). The analyzed period in this study was from 1948 to 2019. The same processes that were applied to the climate model analysis were employed to obtain the ENSO and NPO indices in the observational analysis.
3. SFM efficiency in the CMIP5 models
a. SFM processes in the SFM-Strong and SFM -Weak models
To measure the strength of the NPO–ENSO relationship, SFM efficiency was defined by a 1-yr lagged correlation between the NPO and the subsequent ENSO indices. The SFM efficiencies in the 30 CMIP5 climate models are displayed in Fig. 1. The climate models simulated SFM efficiencies very differently, and these ranged from −0.03 (GISS-E2-H) to 0.47 (NorESM1-M). Half of the climate models showed a significant NPO–ENSO relationship at the 95% confidence level. Weak positive and negative values implied that the SFM did not operate properly in the corresponding models. Overall, the climate models tended to underestimate SFM efficiency when compared to those of the observational results, with the exception of a few models. To understand the diversity among models, the nine highest (top 30%) and nine lowest (bottom 30%) climate models were selected and classified as SFM-Strong and SFM-Weak models, respectively. Then, the characteristics of the SFM in the two groups were compared. We noted that the results in this study were not substantially dependent on the selection of SFM-Strong or SFM-Weak models.
Correlation coefficients between the NPO index and a 1-yr lagged ENSO index. The leftmost bar in blue indicates the observations. Results from the 30 climate models participating in CMIP5 are shown from the second bar to the right in descending order, and the model number is assigned according to the results. Herein, dark green and yellow bars indicate the nine highest (top 30%) and nine lowest (bottom 30%) models, representing SFM-Strong and SFM-Weak models, respectively. Herein, one and two asterisks respectively indicate that the result is significant at the 95% and 99% confidence levels with a Student’s t test.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
To compare the SFM, anomalous SST, SLP, and low-level wind data were regressed onto the normalized NPO index in each climate model. The results were then ensemble-averaged into SFM-Strong or SFM-Weak models (Fig. 2). For the SFM-Strong models (Fig. 2a), NPO-like SLP structure was well observed in winter [December–February (DJF)]), that is, D[0]JF[1], where the 0 and 1 in brackets indicate the zeroth and first years (this convention is used throughout the paper). Its intensity was greater and its location moved farther to the east in the SFM-Strong models compared to those of the SFM-Weak models (Figs. 2b,c). These results were consistent with those of a previous study suggesting an intimate link between SFM efficiency and NPO intensity (Chen and Yu 2020) or eastward-displaced NPO structure (Yeh et al. 2018).
(a) Ensemble of SSTA [(°C); shaded; shading bar is to the right of (b)], anomalous SLP (contours, with 0.25-hPa interval), and anomalous low-level wind (vectors > 0.1 m s−1 at 850 hPa) regressed onto the normalized NPO index in the (a) SFM-Strong models and (b) SFM-Weak models, along with (c) the differences between (a) and (b), where the contour interval of SLPA is 0.125 hPa, for (top) D[0]JF[1], (top middle) MAM[1], (bottom middle) JJA[1], and (bottom) D[1]JF[2]. For the SSTA and wind data, only significant results at greater than the 95% confidence level by a bootstrap method (10 000 times) are shown in the panels.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
In spring (March–May, labeled MAM[1]), the NPO-like SLP anomaly (SLPA) dipole weakened, while southwesterly wind anomalies in the subtropics were well maintained, which further enhanced SST warming in the subtropics, which later expanded farther into the lower latitudes. Note that the patterns of the wind and SST in this season were similar to those of the Pacific meridional mode (PMM; Chiang and Vimont 2004), which is known to be a good precursor for ENSO development associated with the NPO (Chiang and Vimont 2004; Chang et al. 2007; Anderson and Perez 2015; Thomas and Vimont 2016). The NPO/PMM-related westerly anomalies expanded farther toward the equator, which was important for triggering an ENSO event. That is, the anomalous westerly wind generated downwelling oceanic Kelvin waves, which induced SSTA warming from the central to eastern equatorial Pacific. Thus, Bjerknes feedback operated so that an El Niño event could be initiated in summer (July–August, labeled JJA[1]) and developed in the following winter (D[1]JF[2], where the “2” indicates the second year).
As for the SFM-Weak models, the NPO tended to be weakly simulated in winter (D[0]JF[1]; Fig. 2b,c). Accordingly, both anomalous westerly winds and underlying SST warming in the subtropics were weaker in winter until the subsequent spring compared to those of the SFM-Strong models (Fig. 2c). Weak NPO variability was one reason for the weak connection between midlatitudes and tropics in this group. Westerly winds were observed in the western equatorial Pacific in MAM[1], as in the SFM-Strong models, but were weaker in MAM[1], indicating weaker impacts on equatorial wave dynamics. In addition, anomalous easterlies were found in the equatorial eastern Pacific. Easterly winds induced equatorial upwelling, which compensated for the effects of the downwelling Kelvin waves induces by the westerly anomalies in the central Pacific. Therefore, both weaker westerlies in the central Pacific and easterlies in the eastern Pacific constituted a direct reason for weak SFM efficiency.
In Fig. 2, one of the primary differences between the two groups in terms of SFM efficiency was the NPO intensity in winter, which has been noted in previous studies (Chen et al. 2019). To examine the relationship between NPO intensity and SFM efficiency in detail, a scatterplot analysis was conducted (Fig. 3). A significant relationship between NPO intensity and SFM efficiency was found at the 95% confidence level (correlation coefficient of 0.41), regardless of how NPO intensity was defined (refer to section 2). These results imply that stronger NPOs were accompanied by anomalous southwesterly winds around the southern border of the NPO southern lobe, which concurrently led to underlying SST warming due to reduced wind speeds (i.e., WES feedback), supporting the role of NPO intensity in controlling SFM efficiency.
(a) A scatterplot between the standard deviation of NPO and SFM efficiency in the CMIP5 intermodel space, in which the slope of the dotted line shows the regression coefficient. Model numbering denotes the order of SFM efficiency given in Fig. 1. SFM-Strong and SFM-Weak models are identified by green and yellow colors, respectively. The R value at the bottom right indicates the correlation coefficient, and an asterisk indicates the 95% confidence level by Student’s t test. (b) As in (a), but between the intensity of the NPO southern lobe and SFM efficiency, in which the intensity of the NPO southern lobe is defined as the regional average (150°E–150°W, 10°–45°N) of the NPO pattern (second EOF mode of the winter SLP over the North Pacific).
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
In addition to NPO intensity, in order to further understand what was responsible for the SFM diversity in the CMIP5 models, ocean–atmospheric interactions, described as the WES feedback, were investigated. In the subtropics, the strength of ocean–atmosphere interactions could be estimated by the precipitation response. Figure 4 shows the precipitation anomalies associated with the NPO in late winter, from January to April. For the SFM-Strong models, subtropical SSTA gradually increased during this season (DJF to MAM in Fig. 2a). More important, the anomalous cyclonic winds expanded equatorward, leading to low-level convergence (not shown). Although the cyclonic flow and associated positive vorticity existed over a broad region, the precipitation response was stronger in the western Pacific because the climatological ITCZ was located in that region, resulting in strong atmospheric instability.
Ensemble of anomalous precipitation (shaded; mm day−1), SLPA (contours, with 0.25-hPa interval), and anomalous low-level wind (vectors > 0.1 m s−1 at 850 hPa) in January–April regressed onto the normalized NPO index in the (a) SFM-Strong models and (b) the SFM-Weak models, along with (c) the differences between (a) and (b), where only above 95% confidence level by a bootstrap method (10 000 times) is shown for precipitation and winds. The red line in each panel indicates the location of the ensemble mean ITCZ in the Pacific.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
The increased precipitation in the subtropics lay over the Pacific ITCZ in the SFM-Strong models. A Gill-type atmospheric response to anomalous ITCZ precipitation can play a role in enhancing and perpetuating the cyclonic circulation of the southern lobe of the NPO. As a result, the signals of the southern lobe of the NPO could be connected farther toward the tropics. At this moment, the southwesterly winds associated with the NPO southern lobe were likely to transport additional moisture into the subtropics from the tropics, which further activated ocean–atmospheric interactions. However, there were relatively weak precipitation anomalies in the subtropics in the SFM-Weak models. As a result, cyclonic circulation associated with the southern lobe of the NPO was not well developed in the SFM-Weak models when compared with those of the SFM-Strong models. These distinctive differences in the subtropics between the SFM-Strong and SFM-Weak models suggest that they play important roles in the SFM.
So far, we have compared the composites of the SFM-Strong and SFM-Weak models. It was found that the NPO intensity and its relevant atmospheric circulation over the subtropics were stronger in the SFM-Strong models compared to those of the SFM-Weak models. In addition, stronger precipitation responses might be related to stronger atmospheric responses. In the boreal spring, differences in wind anomalies were prominent over the equatorial Pacific. The SFM-Strong models simulated strong westerlies in the western Pacific (Fig. 4a). The westerlies in the western Pacific helped to initiate an El Niño event by developing downwelling Kelvin waves. In the meantime, the SFM-Weak models tended to simulate relatively weak westerlies and easterlies in the western and eastern Pacific, respectively, providing unfavorable conditions for El Niño development. Next, we will show what constitutes these differences.
b. Mean state (ITCZ) differences between SFM-Strong and SFM-Weak models
To understand the causes of the different patterns and evolution of the SFM between the SFM-Strong and SFM-Weak models, we compared the mean states of the two groups. We found distinctive differences between the two groups in terms of mean precipitation, as shown in Fig. 5a. In the SFM-Strong models, the background mean precipitation was more intense in the western Pacific and weaker in the eastern Pacific. In addition, precipitation increased over the most subtropical region of the Pacific, indicating intensification of the northern part of the ITCZ or the poleward shift of the Pacific ITCZ.
(a) Difference in climatological precipitation (shaded; mm day−1) and low-level winds (vectors at 850 hPa) in boreal winter between the SFM-Strong and SFM-Weak models. Contour lines indicate the ensemble mean of climatological precipitation in winter (interval: 2 mm day−1). (b) A scatterplot between Pacific ITCZ latitude (DJF; 120°E–120°W) and SFM efficiency in the CMIP5 intermodel space, where the slope of dotted line shows the regression coefficients. (c) As in (b), but with the meridional precipitation difference (7.5°–10°N minus 5°–7.5°N within 120°E–120°W; the model mean is set to 0). Model numbering denotes the order of SFM efficiency given in Fig. 1. SFM-Strong and SFM-Weak models are identified by green and yellow, respectively. The R values at the bottom right indicate the correlation coefficients, with two asterisks indicating significance at the 99% confidence level by the Student’s t test.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
It was confirmed that the difference in climatological winter precipitation between the SFM-Strong and SFM-Weak models coincided with the difference in the precipitation anomalies associated with the NPO, as shown in Fig. 4c. Specifically, the precipitation difference in Fig. 4c was maximized over the western subtropical Pacific (150°E–180°, 5°–15°N), where the mean precipitation difference was the largest. This result suggests that the stronger precipitation response to the NPO in the SFM-Strong models may arise from favorable background conditions associated with abundant mean precipitation owing to the northward displacement of the ITCZ. In other words, when anomalous low-level convergence occurred over the mean low-level convergence zone, the atmospheric response may be amplified through moisture convergence feedback (Zebiak 1986). Consistently, the wind fields showed stronger convergence over the subtropical North Pacific along the region of increased precipitation.
The above results implied that the climate model with Pacific ITCZs that were shifted poleward tended to have high SFM efficiencies. To address this implication, the relationship between the latitude of the Pacific ITCZ and SFM efficiency was examined in the intermodel space. Herein, the latitude of the ITCZ was defined by averaging the latitudes in which the winter climatological precipitation was maximum at each longitudinal grid point over the area of 120°E–120°W. A scatterplot with ITCZ latitude and SFM efficiency is shown in Fig. 5b. The overall linear relationship verified that the SFM efficiency increased as the Pacific ITCZ moved poleward. The correlation coefficient between these two variables was 0.52 and was significant at the 99% confidence level.
As mentioned above, the northward displacement of the ITCZ seemed to correspond to the intensification of the northern branch of the ITCZ. We defined a meridional difference in the winter mean precipitation around the Pacific ITCZ (7.5°–10°N minus 5°–7.5°N within the region of 120°E–120°W) in each model and compared it with the Pacific ITCZ latitude in the intermodel space. We found that the value of the correlation was 0.89, which was significant at the 99% confidence level. As expected, we also found a significant relationship between the meridional difference in winter mean precipitation and SFM efficiency (Fig. 5c).
c. Ocean–atmospheric interaction associated with ITCZ position and SFM efficiency
Now, we examined whether the climatological position of the ITCZ indeed affected the intensity of air–sea coupling in the subtropical Pacific. To examine the air–sea coupling strength, a subtropical SST index was first defined by the areal average of the SSTA in the region of 120°E–150°W and 5°–15°N in late winter (January–April). Then, anomalous precipitation and low-level winds in early spring were regressed onto the normalized subtropical SST index with a 1-month lag. The 1-month lag was used to avoid the effect of heat flux–induced SST changes and to account for the SST-induced atmospheric response. Figure 6 shows the composite results for the SFM-Strong and SFM-Weak models. As can be seen from the green shading in Figs. 6a and 6b, the response to the positive SSTA showed increased precipitation in the subtropics and weakened trade wind intensity from the equator to the subtropics.
Precipitation and low-level wind anomalies in February–May of the zeroth year (FMAM[0]) regressed onto the normalized SST index (120°E–150°W, 5°–15°N, marked by the outlined rectangle) in January–April of the zeroth year (JFMA[0]) in the (a) SFM-Strong and (b) SFM-Weak models. (c) Their differences, where only above 95% confidence level by a bootstrap method (10 000 times) is plotted.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
However, the detailed structures of the precipitation response were somewhat different between the SFM-Strong models and SFM-Weak Models despite having the same magnitude of SST forcing. First, the overall precipitation pattern was slightly shifted toward the equator in the SFM-Weak models, which may have been related to the location of the ITCZ. Second, the subtropical precipitation response was much stronger in the SFM-Strong models compared to those of the SFM-Weak models. Consistently, the wind responses in the SFM-Strong models were greater to the north compared to those in the south (Fig. 6c).
Upon further examination of the wind responses (Fig. 6c), anomalous westerly winds in the equatorial western Pacific (130°–160°E, 0°–5°N) were greater in the SFM-Strong models (0.75 m s−1) by approximately 50% compared to those of the SFM-Weak models (0.51 m s−1). Stronger westerly winds were expected to effectively force oceanic Kelvin waves while extending WES feedback into the tropics. In addition, the westerly response in the off-equatorial western Pacific was stronger in SFM-Strong models compared to those of SFM-Weak models, implying strong impacts on equatorial ocean dynamics. In addition, the westerly difference was larger off the equator, which induced anticyclonic wind stress curl that contributed to enhancing equatorial recharging (Jin 1997; Li 1997; Anderson et al. 2013; Anderson and Perez 2015). Therefore, off-equatorial westerlies could eventually contribute to triggering an ENSO event.
Note that the elevated atmospheric responses in the SFM-Strong models coincided well with the background mean precipitation differences shown in Fig. 5. This reaffirmed that the intensification of ocean–atmosphere interactions in the subtropics essentially stemmed from enhanced mean precipitation associated with a northward shift of the climatological ITCZ in the Pacific.
We further evaluated the relationship between the ITCZ position and subtropical atmospheric responses (Fig. 7a). The results showed that a poleward-shifted ITCZ was linked to a strong precipitation response in subtropical SST. The value of the correlation was 0.69, which was significant at the 99% confidence level. The strong atmospheric response, in turn, was able to enhance the SSTA by modulating latent heat fluxes at the oceanic surface under conditions of background trade winds in the subtropics. Therefore, the intensity of the precipitation response against changes in SSTA was in line with the strength of WES feedback and was finally linked to SFM efficiency. To demonstrate the above conclusion, the relationship between the atmospheric response to SSTA in the subtropics and SFM efficiency was investigated in the intermodel space as well (Fig. 7b), and a significant relationship between these variables was found at the 99% confidence level (correlation coefficient: 0.67).
(a) Scatterplot between the Pacific ITCZ latitude in winter (x axis, °N) and the precipitation response to the SSTA in the subtropics (130°E–180°, 10°–15°N for January–April) in the CMIP5 intermodel space, in which the slope of the dotted line shows the regression coefficient. As in (a), but (b) between the precipitation response to SSTA in the subtropics and seasonal footprinting mechanism (SFM) efficiency and (c) between Pacific ITCZ latitude and NPO intensity in the CMIP5 intermodel space. SFM-Strong and -Weak models are identified by green and yellow, respectively. The R values at the bottom right indicate the correlation coefficients, with two asterisks indicating significance at the 99% confidence level by a Student’s t test.
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
Meanwhile, Fig. 7c shows a significant correlation coefficient (0.55%; 99% confidence level) between the ITCZ and NPO southern lobe intensity, indicating that the CMIP5 models tended to simulate stronger NPOs when climatological ITCZs were displaced poleward. The above result can also be understood dynamically. As mentioned earlier in section 2a, in regions in which NPO-induced southwesterly winds in the subtropics induced SST warming via reduced evaporation, both surface warming and moisture transport favored positive precipitation. If the ITCZ was located at higher latitudes, stronger precipitation and an associated Gill-type atmospheric response could play a role in intensifying the cyclonic circulation of the southern lobe of the NPO, as shown in Fig. 6, which suggests positive feedback at work. That is, stronger ocean–atmosphere interactions in the subtropics may reinforce NPO intensity, in association with poleward shifts of the background ITCZ. At the same time, such positive feedback among winds, SST, and precipitation acted to maintain the variability of the NPO southern lobe from winter to spring, eventually enhancing the connection to equatorial ocean dynamics.
4. Decadal modulation of SFM efficiency in the observational reanalysis dataset
The CMIP5 model results suggested that the SFM efficiency was largely affected by the climatological position of the Pacific ITCZ, as the northward shift of the ITCZ (i.e., intensification of the northern branch of the ITCZ) strengthened ocean–atmosphere interactions in the subtropics. In the observational reanalysis dataset, on the other hand, it has been previously suggested that SFM efficiency varies by decades (Yeh et al. 2018). This earlier work attributed the decadal variation in SFM efficiency to decadal changes in the spatial structure of the NPO. In the present study, however, based on the findings from the climate models, the decadal characteristics of SFM efficiency were examined in terms of the tropical mean state associated with the Pacific ITCZ.
Figure 8a shows the 15-yr moving correlation coefficients between the NPO and ENSO indices (black), which represent the decadal modulation of SFM efficiency in the observational dataset. In addition, an SFM frequency, defined as the frequencies of SFM occurrence within a 15-yr moving period, is also illustrated (blue). The SFM occurrence was counted when a positive (negative) NPO was followed by an El Niño (La Niña) event one year later. We noted that the results did not depend on the criteria to identify NPO/ENSO events. To clarify this point, three criteria were considered, wherein NPO/ENSO occurrence was defined according to when their normalized indices exceeded ±0.5, ±0.75, and ±1.0, respectively. Regardless of these criteria, the overall trends of SFM frequency were similar to the 15-yr moving NPO–ENSO correlation. This result verified that the NPO–ENSO correlation (i.e., SFM efficiency) properly reflected the observed SFM frequency on decadal time scales, which is consistent with the results of Chen et al. (2019; refer to Fig. 7 therein).
(a) The 15-yr moving correlation coefficients between the NPO index and the ENSO index with 1-yr lag (left axis; black line with circles). Closed circles indicate significance at the 95% confidence level with a Student’s t test. SFM frequency is illustrated by blue lines (right axis; number of SFM occurrences per 15 years). For the blue lines, times signs, open squares, and plus signs respectively indicate the SFM frequency according to the ±0.5, ±0.75, and ±1.0 standard deviation of the NPO/ENSO index. (b) The 15-yr moving winter ITCZ latitude over 150°E–120°W (right axis; °N; indicated with the green line with open squares) and the 15-yr low-pass filtered ITCZ latitude (°N; green solid line). Also shown is the meridional difference (7.5°–10°N minus 5°–7.5°N over 150°E–120°W) of climatological precipitation in winter (normalized; purple line with plus signs). (c) The 15-yr moving standard deviation of the NPO index (right axis, red line). In (b) and (c), the black line is the same as in (a). In (a)–(c), light-blue- and light-red-shaded periods respectively indicate SFM high- and low-efficiency periods, judged by results of (a), wherein the first and the second light-blue-shaded periods are P1 (1957–71) and P3 (1992–2005), respectively, and the light-red-shaded period is P2 (1972–86).
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
The correlation coefficient between the NPO and ENSO indices during 1948–2019 was 0.37 (Fig. 1a), which was significant at the 95% confidence level with a two-tailed Student’s t test. As shown in Fig. 8a, however, their relationship was not stationary and exhibited prominent decadal modulation. The highest and lowest correlation coefficients were 0.70 and −0.08, respectively. From a linear perspective, this result indicated that the preceding NPO explained ENSO variability up to ~50% in one period, whereas it did not explain any variability at all in the other period. The observed SFM efficiency was roughly high during 1957–71 and 1992–2005, but low during 1972–86. In this study, the average of the two high-correlation periods (i.e., 1957–71 and 1992–2005) was considered to be an SFM high-efficiency period, whereas 1972–86 was considered to be a SFM low-efficiency period; these are counterparts of the SFM-Strong and SFM-Weak models in CMIP5 results, respectively.
Considering the results of the CMIP5 models, we also examined decadal changes in the tropics and NPO intensity in the observational dataset. Before elucidating pattern changes, we compared temporal variations in the latitude of the climatological ITCZ and NPO intensity, which exhibited pronounced decadal variations. Figure 8b presents the 15-yr moving mean of the latitude of the winter ITCZ, defined by the same method as in the CMIP5 models (green lines). It was determined that the observed winter ITCZ had migrated between 7.5° and 8.2°N in the meridional direction during the past decades. Herein, the northward (southward) migration largely corresponded to the intensification of the northern (southern) branch of the ITCZ, as indicated by the meridional difference in climatological winter precipitation (purple line). Interestingly, the meridional migrations of the ITCZ coincided with changes in SFM efficiency, supporting the conclusion that SFM efficiency is also closely related to the meridional ITCZ position in the observational dataset. Likewise, decadal variation in NPO intensity was traced for comparison with SFM efficiency (Fig. 8c). The NPO intensity, assessed by its standard deviation within 15-yr moving windows, showed a phase similar to that of SFM efficiency.
The above results show good agreement between the observations and intermodel comparison. The spatial features of the observational results were also consistent with the CMIP5 results. Figure 9 presents anomalous SST, SLP, and low-level winds regressed onto the observed NPO index in both the SFM high-efficiency and SFM low-efficiency periods, in the same manner as that of the model analysis. Note that the NPO index was newly defined in each period by removing the mean and trend of the corresponding period. In the SFM high-efficiency period (Fig. 9a), an NPO-like SLP structure was clearly observed in D[0]JF[1], and along the southern border, anomalous westerly winds were located. Such westerly winds started WES feedback. Accordingly, the westerly wind anomalies and SST warming continued to grow in the central equatorial Pacific in MAM[1], even after the NPO-like SLP structure disappeared. In JJA[1], both westerly winds and SST warming intensified, leading to El Niño events in the following winter (D[1]JF[2]).
SSTA (shaded; °C), SLPA (contours, 0.5-hPa interval), and anomalous low-level wind (vectors > 0.5 at 850 hPa; m s−1) regressed onto the normalized NPO index in (a) the SFM high-efficiency periods (P1: 1957–71 and P3: 1992–2005) and (b) the SFM low-efficiency period (P2: 1972–86) for (top) D[0]JF[1], (top middle) MAM[1], (bottom middle) JJA[1], and (bottom) D[1]JF[2]. (c) The difference of (a) minus (b), where SSTA and low-level winds are plotted when the difference between the two periods (P3 − P2 or P1 − P2) was significant at the 95% confidence level by a bootstrap method (10 000 times).
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
In contrast, in the SFM low-efficiency period (Fig. 9b), the NPO-like SLP dipole structure was less clear and appeared to have a single cell structure. Notably, the negative SLPA associated with the NPO southern lobe did not expand toward the tropics, thus westerly wind anomalies did not reach the deep tropics (Figs. 9b,c). From D[0]JF[1] to MAM[1], considerable SST warming was observed in the subtropical eastern North Pacific (120°–150°W, 15°–30°N). However, this was confined within the subtropics, and was not connected to the equatorial tropics by westerlies. Subtropical SST warming decayed in JJA[1]. Simultaneously, weak SST warming was found to grow in the central equatorial Pacific, accompanied by anomalous easterly winds in the eastern equatorial Pacific. This acted to restrain El Niño development.
Figure 10 presents anomalous precipitation, SLP, and low-level winds in late winter, regressed onto the NPO index during SFM high- and low-efficiency periods. For the SFM high-efficiency period (Fig. 10a), a meridional SLP dipole was clearly observed from late winter to early spring. Along the southern area of the NPO southern lobe, anomalous westerly winds dominated, spreading from the midlatitudes to the tropics. Notably, a positive precipitation anomaly was formed in the western tropical Pacific along the Pacific ITCZ. Meanwhile, for the SFM low-efficiency period (Fig. 10b), there appeared to be a negative SLP monopole, rather than a meridional SLP dipole, which indicated that, overall, the NPO southern lobe was weakened in the subtropics to the point that it became disconnected from the tropics. With a weakened NPO southern lobe, both anomalous westerly winds and positive precipitation were no longer observed in the subtropics. Consequently, one of the distinct features during the SFM high-efficiency periods was strengthened precipitation in the subtropics along the ITCZ region, as denoted by the shading in Fig. 10c.
The NPO index–based regression maps of anomalous precipitation (shaded; mm day−1), SLPA (contours, at 0.5-hPa interval), anomalous low-level winds (vectors at 850 hPa; m s−1) from January to April during (a) SFM high-efficiency and (b) SFM low-efficiency periods. (c) The differences between (a) and (b), whereprecipitation and low-level winds are marked when the difference between two periods (P3 − P2 or P1 − P2) was significant at the 95% confidence level by a bootstrap method (10 000 times).
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
Next, we also examined winter-mean precipitation differences between the SFM high- and SFM low-efficiency periods. In Fig. 11, the climatological ITCZ in the Pacific was located off the northern side of the equator, parallel to 7°–8°N. During the SFM-high efficiency period, precipitation seemed to increase to the north of the ITCZ rather than in the southern region, as indicated by Fig. 8b (and vice versa). This feature agreed with the results from the CMIP5 analysis (Fig. 8b), in which precipitation over the northern (southern) branch of the Pacific ITCZ was enhanced in the SFM-Strong (SFM-Weak) models. Additionally, we noted that the strong precipitation anomalies in the subtropical North Pacific forced by the NPO were found around the region with the largest increase in mean background precipitation (150°E–150°W; Figs. 10a and 11). This implied that intensified ocean–atmosphere interactions during the SFM high-efficiency periods were attributable to the increased mean precipitation in the northern branch of the Pacific ITCZ.
Precipitation difference in winter between the SFM high- and low-efficiency periods (shaded; mm day−1). The climatology of precipitation is illustrated by contour lines as a reference, where values greater than 4 mm day−1 are shown. Hatched lines indicate that the difference was significant at the 95% confidence level by a bootstrap method (10 000 times).
Citation: Journal of Climate 34, 13; 10.1175/JCLI-D-20-0809.1
5. Discussion and summary
In this study, we attempted to identify the factors determining SFM efficiency by analyzing datasets from both CMIP5 models and reanalysis of observational data. In particular, we focused on the position of the ITCZ in winter and its association with the surrounding distribution of climatological precipitation when the SFM was vigorous. It was found that climate models with poleward-shifted Pacific ITCZ climatology tended to have strong SFM efficiency. Consistently for the observational data reanalysis, it was revealed that the SFM efficiency increased during the decades in which the Pacific ITCZ extended poleward.
Previous studies have argued that NPO characteristics are important to SFM efficiency. Likewise, we also found a significant relationship between NPO intensity and SFM efficiency. However, in this study, we assigned a greater priority to ITCZ position compared to that of NPO intensity, based on the role of the ITCZ in modulating NPO characteristics, as discussed above. Further, the greater degree of correlation between ITCZ latitude and SFM efficiency (0.52) relative to that of NPO intensity and SFM efficiency (0.41) further supports our conclusions.
The influence of the position of the climatological ITCZ in winter on SFM efficiency manifested itself through two processes. The first was associated with WES feedback, a key mechanism for the SFM. The ITCZ is known as the region in which the ocean and atmosphere are strongly coupled due to strong atmospheric instability. This environment enables a perturbation to grow readily through vigorous ocean–atmosphere interactions. The northward shift in the climatological position of the ITCZ facilitated the amplification of atmospheric and oceanic anomalies triggered by the NPO. Namely, southern circulation anomalies of the NPO, which sustain the SST anomalies in the subtropical ocean, were further reinforced by strengthened convection anomalies due to WES feedback and an ensuing Gill-type response. This intensified tropical–subtropical coupling and led to increased SFM efficiency.
Another role of the climatological ITCZ appeared rather indirectly. We showed that NPO variability tended to increase as the mean ITCZ position moved poleward (Figs. 7 and 8). This may not only be because of the intense ocean–atmosphere interactions in the subtropics, but also because of the tropical mean state that corresponded to the position of the ITCZ. In this light, recent studies have suggested a possible mechanism by examining the energetics of the NPO (Tanaka et al. 2016; Sung et al. 2019, 2020). They have shown that the growth and maintenance of the NPO depend on the background temperature gradient over the North Pacific, as the extraction of available potential energy from the basic state represents a major process required to maintain the NPO. Through a Rossby wave bridge, changes in the tropical mean state have been found to modulate the thermal structure over the North Pacific and thus NPO characteristics as well (Sung et al. 2019). The SFM-Strong and SFM-Weak CMIP5 models showed roughly consistent differences in the tropical mean state compared to those of the observational data reanalysis, namely, cooler SST was found over the eastern equatorial Pacific in the SFM-Strong models compared to those of the SFM-Weak models, although the signals were much weaker due to the large diversity among model mean states (not shown). This may suggest that a similar modulation effect by the tropical mean state in the SFM-Strong models also acts on the background temperature gradient over the North Pacific to strengthen NPO intensity.
It would be interesting to discuss what controls the latitudinal position of the Pacific ITCZ. On an interannual time scale, it is known that ENSO plays the most dominant role in determining the position of the ITCZ (Rasmusson and Carpenter 1982; Vecchi and Harrison 2006; Schneider et al. 2014). On a decadal time scale, it seems that the Pacific meridional mode or North Pacific Gyre Oscillation are associated with the ITCZ position in the observational data reanalysis (not shown).
Recently, distinct features of midlatitude and ENSO linkage have been highlighted with regard to ENSO diversity (Yu and Fang 2018; Wu et al. 2019; Park et al. 2020). For example, Yu and Fang (2018) argued that the SFM is a critical source of ENSO complexity because it makes the evolution of an ENSO more episodic and irregular by involving subtropical forcing. Park et al. (2020) suggested that La Niña events with midlatitude linkage tend to have a long persistency (more than two years) with mega-ENSO patterns (Wang et al. 2013). Given the results from previous studies, the decadal condition of the Pacific ITCZ may also serve as an indicator of ENSO complexity, since the ITCZ is closely associated with SFM efficiency. We believe that further relevant studies are valuable to pursue these issues.
Acknowledgments
We appreciate the constructive comments and suggestions provided by the three 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 Korea government (MSIT) (NRF-2020R1C1C1006569; NRF-2018R1A5A1024958).
Data availability statement
The SST dataset (Extended Reconstructed SST v5) and atmospheric dataset (National Centers for Environmental Prediction–National Center for Atmospheric Research) were downloaded from NOAA (https://www.ncei.noaa.gov/pub/data/cmb/ersst/v5/netcdf/.) In this study, historical simulation datasets that participated in phase 5 of the Coupled Model Intercomparison Project (CMIP5) were downloaded CMIP5 (https://esgf-node.llnl.gov/projects/cmip5).
REFERENCES
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, 2885–2901, https://doi.org/10.1175/2010JCLI3205.1.
Amaya, D. J., 2019: The Pacific Meridional Mode and ENSO: A review. Curr. Climate Change Rep., 5, 296–307, https://doi.org/10.1007/s40641-019-00142-x.
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, 7643–7661, https://doi.org/10.1175/JCLI-D-19-0040.1.
Anderson, B. T., and R. C. Perez, 2015: ENSO and non-ENSO induced charging and discharging of the equatorial Pacific. Climate Dyn., 45, 2309–2327, https://doi.org/10.1007/s00382-015-2472-x.
Anderson, B. T., R. C. Perez, and A. Karspeck, 2013: Triggering of El Niño onset through trade wind-induced charging of the equatorial Pacific. Geophys. Res. Lett., 40, 1212–1216, https://doi.org/10.1002/grl.50200.
Baxter, S., and S. Nigam, 2015: Key role of the North Pacific oscillation–west Pacific pattern in generating the extreme 2013/14 North American winter. J. Climate, 28, 8109–8117, https://doi.org/10.1175/JCLI-D-14-00726.1.
Bond, N. A., J. E. Overland, M. Spillane, and P. Stabeno, 2003: Recent shifts in the state of the North Pacific. Geophys. Res. Lett., 30, 2183, https://doi.org/10.1029/2003GL018597.
Ceballos, L. I., E. Di Lorenzo, C. D. Hoyos, N. Schneider, and B. Taguchi, 2009: North Pacific Gyre Oscillation synchronizes climate fluctuations in the eastern and western boundary systems. J. Climate, 22, 5163–5174, https://doi.org/10.1175/2009JCLI2848.1.
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.
Chen, S., and R. Wu, 2018: Impacts of winter NPO on subsequent winter ENSO: Sensitivity to the definition of NPO index. Climate Dyn., 50, 375–389, https://doi.org/10.1007/s00382-017-3615-z.
Chen, S., and B. Yu, 2020: Projection of winter NPO-following winter ENSO connection in a warming climate: Uncertainty due to internal climate variability. Climatic Change, 162, 723–740, https://doi.org/10.1007/s10584-020-02778-3.
Chen, S., W. Chen, B. Yu, and H.-F. Graf, 2013: Modulation of the seasonal footprinting mechanism by the boreal spring Arctic Oscillation. Geophys. Res. Lett., 40, 6384–6389, https://doi.org/10.1002/2013GL058628.
Chen, S., B. Yu, and W. Chen, 2014: An analysis of the physical process of the influence of AO on ENSO. Climate Dyn., 42, 973–989, https://doi.org/10.1007/s00382-012-1654-z.
Chen, S., L. Y. Song, and W. Chen, 2019: Interdecadal modulation of AMO on the winter North Pacific oscillation following winter ENSO relationship. Adv. Atmos. Sci., 36, 1393–1403, https://doi.org/10.1007/s00376-019-9090-1.
Chiang, J. C. H., and D. J. Vimont, 2004: Analogous Pacific and Atlantic meridional modes of atmosphere–ocean variability. J. Climate, 17, 4143–4158, https://doi.org/10.1175/JCLI4953.1.
Di Lorenzo, E., and Coauthors, 2008: North Pacific Gyre Oscillation links ocean climate and ecosystem change. Geophys. Res. Lett., 35, L08607, https://doi.org/10.1029/2007GL032838.
Hartmann, D. L., 2015: Pacific sea surface temperature and the winter of 2014. Geophys. Res. Lett., 42, 1894–1902, https://doi.org/10.1002/2015GL063083.
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.
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.
Kistler, R., and Coauthors, 2001: The NCEP-NCAR 50-Year Reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82, 247–267, https://doi.org/10.1175/1520-0477(2001)082<0247:TNNYRM>2.3.CO;2.
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.
Linkin, M. E., and S. Nigam, 2008: The North Pacific Oscillation–west Pacific teleconnection pattern: Mature-phase structure and winter impacts. J. Climate, 21, 1979–1997, https://doi.org/10.1175/2007JCLI2048.1.
Park, J.-H., S.-I. An, J.-S. Kug, Y.-M. Yang, T. Li, and H.-S. Jo, 2020: Mid-latitude leading double-dip La Niña. Int. J. Climatol., 41, E1353–E1370, https://doi.org/10.1002/JOC.6772.
Park, J.-Y., S.-W. Yeh, J.-S. Kug, and J. Yoon, 2013: Favorable connections between seasonal footprinting mechanism and El Niño. Climate Dyn., 40, 1169–1181, https://doi.org/10.1007/s00382-012-1477-y.
Rasmusson, E. M., and T. H. Carpenter, 1982: Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño. Mon. Wea. Rev., 110, 354–384, https://doi.org/10.1175/1520-0493(1982)110<0354:VITSST>2.0.CO;2.
Rogers, J. C., 1981: The North Pacific Oscillation. J. Climatol., 1, 39–57, https://doi.org/10.1002/joc.3370010106.
Schneider, T., T. Bischoff, and G. H. Haug, 2014: Migrations and dynamics of the intertropical convergence zone. Nature, 513, 45–53, https://doi.org/10.1038/nature13636.
Screen, J. A., and I. Simmonds, 2014: Amplified mid-latitude planetary waves favour particular regional weather extremes. Nat. Climate Change, 4, 704–709, https://doi.org/10.1038/nclimate2271.
Song, L., Y. Li, and W. Duan, 2016: The influence of boreal winter extratropical North Pacific Oscillation on Australian spring rainfall. Climate Dyn., 47, 1181–1196, https://doi.org/10.1007/s00382-015-2895-4.
Sung, M.-K., H.-Y. Jang, B.-M. Kim, S.-W. Yeh, Y.-S. Choi, and C. Yoo, 2019: Tropical influence on the North Pacific Oscillation drives winter extremes in North America. Nat. Climate Change, 9, 413–418, https://doi.org/10.1038/s41558-019-0461-5.
Sung, M.-K., C. Yoo, S.-W. Yeh, Y. Kosaka, and S.-I. An, 2020: Characteristics of the North Pacific Oscillation in CMIP5 models in relation to atmospheric mean states. J. Climate, 33, 3809–3825, https://doi.org/10.1175/JCLI-D-19-0446.1.
Tanaka, S., K. Nishii, and H. Nakamura, 2016: Vertical structure and energetics of the western Pacific teleconnection pattern. J. Climate, 29, 6597–6616, https://doi.org/10.1175/JCLI-D-15-0549.1.
Thomas, E. E., and D. J. Vimont, 2016: Modeling the mechanisms of linear and nonlinear ENSO responses to the Pacific meridional mode. J. Climate, 29, 8745–8761, https://doi.org/10.1175/JCLI-D-16-0090.1.
Vecchi, G. A., and D. E. Harrison, 2006: The termination of the 1997–98 El Niño. Part I: Mechanisms of oceanic change. J. Climate, 19, 2633–2646, https://doi.org/10.1175/JCLI3776.1.
Vimont, D. J., 2010: Transient growth of thermodynamically coupled variations in the tropics under an equatorially symmetric mean state. J. Climate, 23, 5771–5789, https://doi.org/10.1175/2010JCLI3532.1.
Vimont, D. J., D. S. Battisti, and A. C. Hirst, 2001: Footprinting: A seasonal connection between the tropics and mid-latitudes. Geophys. Res. Lett., 28, 3923–3926, https://doi.org/10.1029/2001GL013435.
Vimont, D. J., D. S. Battisti, and A. C. Hirst, 2003a: The seasonal footprinting mechanism in the CSIRO general circulation models. J. Climate, 16, 2653–2667, https://doi.org/10.1175/1520-0442(2003)016<2653:TSFMIT>2.0.CO;2.
Vimont, D. J., J. M. Wallace, and D. S. Battisti, 2003b: The seasonal footprinting mechanism in the Pacific: Implications for ENSO. J. Climate, 16, 2668–2675, https://doi.org/10.1175/1520-0442(2003)016<2668:TSFMIT>2.0.CO;2.
Vimont, D. J., M. Alexander, and A. Fontaine, 2009: Midlatitude excitation of tropical variability in the Pacific: The role of thermodynamic coupling and seasonality. J. Climate, 22, 518–534, https://doi.org/10.1175/2008JCLI2220.1.
Walker, G. T., and E. W. Bliss, 1932: World weather V. Memoirs Roy. Meteor. Soc., 4, 53–84.
Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109, 784–812, https://doi.org/10.1175/1520-0493(1981)109<0784:TITGHF>2.0.CO;2.
Wang, B., J. Liu, H. J. Kim, P. J. Webster, S. Y. Yim, and B. Xiang, 2013: Northern Hemisphere summer monsoon intensified by mega-El Niño/Southern Oscillation and Atlantic multidecadal oscillation. Proc. Natl. Acad. Sci. USA, 110, 5347–5352, https://doi.org/10.1073/pnas.1219405110.
Wang, L., W. Chen, and R. Huang, 2007: Changes in the variability of North Pacific Oscillation around 1975/1976 and its relationship with East Asian winter climate. J. Geophys. Res., 112, D11110, https://doi.org/10.1029/2006JD008054.
Wu, X., Y. M. Okumura, and P. N. DiNezio, 2019: What controls the duration of El Niño and La Niña events? J. Climate, 32, 5941–5965, https://doi.org/10.1175/JCLI-D-18-0681.1.
Xie, S.-P., and S. G. H. Philander, 1994: A coupled ocean–atmosphere model of relevance to the ITCZ in the eastern Pacific. Tellus, 46A, 340–350, https://doi.org/10.3402/tellusa.v46i4.15484.
Yeh, S.-W., D.-W. Yi, M.-K. Sung, and Y. H. Kim, 2018: An eastward shift of the North Pacific Oscillation after the mid-1990s and its relationship with ENSO. Geophys. Res. Lett., 45, 6654–6660, https://doi.org/10.1029/2018GL078671.
Yu, J.-Y., and S.-W. Fang, 2018: The distinct contributions of the seasonal footprinting and charged–discharged mechanisms to ENSO complexity. Geophys. Res. Lett., 45, 6611–6618, https://doi.org/10.1029/2018GL077664.
Zebiak, S. E., 1986: Atmospheric convergence feedback in a simple model for El Nino. Mon. Wea. Rev., 114, 1263–1271, https://doi.org/10.1175/1520-0493(1986)114<1263:ACFIAS>2.0.CO;2.
Zhao, J., J.-S. Kug, J.-H. Park, and S.-I. An, 2020: Diversity of North Pacific Meridional Mode and its distinct impacts on El Niño–Southern Oscillation. Geophys. Res. Lett., 47, e2020GL088993, https://doi.org/10.1029/2020GL088993.
