Does the onset of the subtropical jet stream associated with the North Atlantic contribute to ENSO-like sea surface temperature (SST) anomalies? Why do low-level westerly winds over the equator strengthen? (850 hPa) and contribute to the development of El Niño-like sea surface temperature anomalies?

El Niño‐Southern Oscillation (ENSO) is the dominant atmosphere‒ocean coupling system over equatorial Pacific and has substantial influences on global climate/weather. Here, we find that the west‒east sea surface temperature (SST) gradient over the northeastern Atlantic can influence the remote development of ENSO through the wave train along the subtropical westerly jet in boreal winter. Specifically, a positive SST gradient favors a positive precipitation anomaly over the northeastern Atlantic by inducing a cyclonic anomaly. Diabatic heating induced by anomalous rainfall excites an anticyclonic response in the upper troposphere, which propagates eastward along the subtropical westerly jet and eventually leads to an anticyclonic anomaly over the eastern tropical Pacific. This anticyclonic anomaly tends to cause an anomalous divergence there, which promotes atmospheric convection. As a result, the low‐level westerly winds over the equator strengthen (850 hPa), contributing to the development of an El Niño‐like SST anomaly. In this study, we find that the North Atlantic SST anomaly modulates the ENSO‐like SST anomaly in the eastern tropical Pacific via the Rossby wave train along the subtropical westerly jet during boreal winter, accounting for 34%–48% of ENSO‐related SST anomaly. The further results show that a positive (negative) SST gradient in the northeast Atlantic is conducive to a positive (negative) precipitation anomaly here, which serves as a heating (cooling) exciting an atmospheric Rossby wave propagated eastward along subtropical westerly jet. The divergence (convergence) related to this wave train supports the development of El Niño (La Niña)‐like SST. The findings in this paper about the link between the North Atlantic and ENSO through the Rossby wave train along the subtropical westerly jet may be important for understanding the development of ENSO and its phase locking (peak during winter).

Figure 1. Regression map of geopotential height (shading; unit: m) and wave activity flux (Text S1 in Supporting Information S1; vectors; unit: m2 s− 2) at 250 hPa against the time series (PC2) of the empirical orthogonal function in An, Chen, et al. (2022). The white dots indicate that the 250‐hPa geopotential height signal is significant at the 90% confidence level using the Student's t‐test. This figure is replotted by using the data and method in An, Chen, et al. (2022). A, B, C, …, J denote the central positions of the wave train.

1. Introduction

El Niño‐Southern Oscillation (ENSO), the most significant climate phenomenon operating on interannual timescale, was considered to be a result of the atmosphere‒ocean interaction (Bjerknes, 1969), which has profound influences on global climate and weather (McPhaden et al., 2006). Hence, various efforts have been made to understand the ENSO developing process from observational analysis (e.g., Bjerknes, 1969; Philander, 1983), theory (e.g., Jin, 1997; Suarez & Schopf, 1988; Wytrki, 1975), and numerical model (e.g., Chen et al., 2004; Tang et al., 2018; Zebiak & Cane, 1987). Recently, Cai et al. (2019) and C. Z. Wang (2019) demonstrated that a better understanding of the global teleconnections or inter‐ocean interactions is necessary to improve ENSO forecast. As for the interbasin interaction, many previous studies have suggested that the variations in the Atlantic may affect ENSO (e.g., Jansen et al., 2009; Park et al., 2022; Rong et al., 2010; C. Wang, 2006; C. Wang et al., 2010; X. Wang et al., 2011; Yu et al., 2015), with most influences exhibiting a lag of several months prior to ENSO events. Research in to the roleof the Atlanticon ENSO hasma in lyfocused on the tropical Atlantic (e.g.,Jansenetal.,2009; Rong et al., 2010; C. Wang, 2006). However, few studies focused on the North Atlantic sea surface temperature (SST), in particular the linkage between the North Atlantic and ENSO through the atmospheric Rossby wave train.

Barnston and Livezey (1987) first found that the linkage of the North Atlantic SST with El Niño is through atmospheric teleconnection pattern which is similar to the East Atlantic/West Russia teleconnection. Specifically, the cold SST anomaly in the North Atlantic produces a Rossby wave‐like train that strengthens the Siberian high and the associated East Asian winter monsoon. These changes are followed by an anomalous cyclone in the extratropical western Pacific during the subsequent spring, which favors the occurrence of the westerly wind bursts, consequently triggering an El Niño event. X. Wang et al. (2013) further confirmed the remote influence of the North Atlantic SST on the El Niño event based on atmospheric general circulation model experiments. These studies emphasized the role of the Siberian High in winter, which is related to the wave train along high‐latitudes of the Northern Hemisphere (i.e., East Atlantic/West Russia teleconnection). However, the wave train along the subtropical westerly jet (hereafter termed the subtropical wave train (SWT)) is also an important pattern in boreal winter, exhibiting pronounced interannual variability (An et al., 2020, An, Chen et al., 2022, An, Sheng et al., 2022; Hoskins & Ambrizzi, 1993; Huang et al., 2020; Li & Sun, 2015). The SWT is a Rossby wave‐like train with alternating centers along the subtropical westerly jet over Eurasia, which can be obtained by the second empirical orthogonal function in the domain of (80°W− 170°E, 5− 80°N) during November to January (NDJ) (An, Chen, et al., 2022). It accounts for 13.8% of the total variance, is significant based on the criterion by North et al. (1982) and not sensitive to the domain (An, Chen, et al., 2022). According to An, Chen, et al. (2022), the SWT may contribute to ENSO development by causing an anticyclonic anomaly in the upper troposphere of the eastern Pacific (Figure 1). Unfortunately, underlying processes about the possible relationships between the SWT and ENSO have not been revealed. Furthermore, previous studies (e.g., An, Chen, et al., 2022; Huang et al., 2020; Jiao et al., 2019; Song et al., 2014) suggest that the SWT as shown in Figure 1 may be related to the North Atlantic variations. These hypotheses imply that the SWT may connect ENSO to the North Atlantic variations.

As mentioned above, although the SWT has been proposed as a possible bridge linking ENSO to the North Atlantic, the underlying physical mechanisms remain unclear and are not well constrained in current studies (An, Chen, et al., 2022). In this context, we will first attempt to examine whether the SWT really supports the development of ENSO. We will then discuss the possible impact of the SST anomaly in the North Atlantic on the SWT. The remainder of this study is organized as follows: Section 2 presents the data and methods used in this paper, Section 3 explores the role of the SWT on ENSO, Section 4 addresses the factors such as diabatic heating responsible for the SWT, and Section 5 provides a summary and discussion.

Data and Methods

2.1. Data

The observational reanalysis data (including monthly and daily data) is National Centers for Environmental Prediction (NCEP)/U.S. Department of Energy Reanalysis II, which is assimilated by a state‐of‐the‐art analysis/ forecast system using past data from 1979 to the present (Kanamitsu et al., 2002). It has a horizontal resolution of 2.5° × 2.5° and 17 pressure levels from 1,000 to 10 hPa. The global monthly SST data set is the Extended Reconstruction Sea Surface Temperature version 5 (ERSSTv5) with a horizontal resolution of 2.0° × 2.0°, which is derived from the International Comprehensive Ocean–Atmosphere Data set (ICOADS) available from 1854 to the present (Huang et al., 2017). The monthly precipitation data set is the Global Precipitation Climatology Project (GPCP), which combines observations and satellite precipitation data into 2.5° × 2.5° global grids from 1979‐present (Adler et al., 2018).

2.2. Methods

To quantify the diabatic heating released by rainfall, we calculated q1 (K day− 1) according to Yanai et al. (1973) using high‐frequency (i.e., daily) data:

∂T

Q1 = cp📷 − cp(ωσ − V · ∇T) = cpq1, (1)

∂t

where Q1 is the total diabatic heating (i.e., apparent heat source). Here, cp denotes the specific heat at constant pressure, T the temperature, t the time, p the pressure, ω the vertical velocity at pressure levels, σ the static stability, V the horizontal wind vector, ∇ the horizontal gradient operator.

The linear baroclinic model (LBM; Watanabe & Kimoto, 2000) has been widely used to examine the atmospheric response to prescribed diabatic heating (e.g., An, Sheng, et al., 2022; Luo et al., 2022; Watanabe & Kimoto, 2000). The LBM is a powerful tool to help understanding the complicated feedbacks in the dynamical atmosphere by removing nonlinear processes (Watanabe & Kimoto, 2000). As a result, dynamic framework of the LBM is simplified so that physical processes can be interpreted much more easily. In this study, the LBM was run up 30 days during NDJ based on a background field provided by NCEP reanalysis data. The average of the outputs from day 1–20 was further analyzed as a stable response.

In addition, the linear regression analysis was used to investigate the connection between different variables. A long‐time linear trend was removed for all data used in this study. The anomaly was calculated based on a climatological state of 1981–2010.

Role of the Subtropical Wave Train Along the Subtropical Westerly Jet on the ENSO‐Like SST As seen in Figure 1, the SWT tends to bifurcate into two branches near the dateline. According to the previous, a strong absolute vorticity gradient can be served as a waveguide that supports the propagation of the Rossby wave (Hoskins & Ambrizzi, 1993; Müller & Ambrizzi, 2007). In this study, we find that the strong absolute vorticity gradient associated with the SWT decreases obviously in the east of the dateline in the North Pacific, which leads to a Rossby wave reflection (Figure 1 and Figure S1a in Supporting Information S1). In addition, the distribution of the stationary Rossby wavenumber (i.e., 📷 around the subtropical westerly jet) also supports such a conclusion (Text S2 and Figure S1b in Supporting Information S1). As a result, the SWT results in a positive geopotential height anomaly in the upper troposphere of the eastern Pacific, which is consistent with the development of an ENSO‐like SST. To further illustrate that the SWT is independent of ENSO, we have checked the SWT after removing the ENSO signal (i.e., Niño3.4 index) and find that the SWT still appears even without ENSO (Figure S2 in Supporting Information S1). To rule out the possibility that the anomalous anticyclone over the subtropical eastern Pacific influences the SWT, we redefined a SWT index (Text S3 in Supporting Information S1). The new index primarily captures the wave train centers indicated by the seven purple boxes in Figure S3a in Supporting Information S1, which are unrelated to ENSO (Figure S2 in Supporting Information S1). The correlation coefficient between the newly defined index and the original PC2 index used in the manuscript is 0.91 (significant at the 0.01 level), indicating that the new index effectively represents the SWT. Using this new index, we regressed the 200‐hPa geopotential height field and found that a significant anomalous anticyclone center still appears over the subtropical eastern Pacific (Figure S3b in Supporting Information S1). The associated wave activity flux originates from the upstream cyclonic anomaly, closely resembling the SWT pattern shown in Figure 1. This further supports the conclusion that the anomalous anticyclone over the subtropical eastern Pacific is a result of the SWT, rather than an independent or ENSO‐driven feature. That is, the SST anomalies over the eastern tropical Pacific during NDJ may be linked to the SWT, as evidenced by a correlation coefficient of 0.57 between the new SWT index and the area‐averaged SST anomalies over the region (5°S–5°N, 170–80°W). In addition, from the scatterplot of the PC2 and the SST anomaly in the domain (5°S–5°N, 170–80°W), the probability that the PC2 and the SST anomaly are in the same phase (i.e., 22.5% (30%) for the positive (negative) phase) are remarkably larger than that in the opposite phase (i.e., 7.5% and 15%) (Figure 2). These statistics

📷

Figure 2. Scatterplot between the PC2 and sea surface temperature (SST) anomaly in the domain (5°S–5°N, 170–80°W) marked by the box in Panel (3c). The model numbering denotes years from 1979 (i.e., 79) to 2019 (i.e., 19). The blue circle means the PC2 and the SST index are more than 0.8 or less than − 0.8.

further indicate that the positive phase of the SWT tends to warm the SST in the tropical eastern Pacific, and vice versa. These conclusions are further supported by the distribution of the SST index when lagged behind PC2 (Figure S4 in Supporting Information S1).

Next, we examine how the SWT modulates the ENSO‐like SST anomaly. Figure 3a shows the SWT‐related divergent wind and velocity potential at 250 hPa in NDJ. The SWT generically exhibits a series of alternating convergence and divergence centers along the subtropical westerly jet. Clearly, there is a strong center of divergence in the upper troposphere of the tropical eastern Pacific (Figure 3a), which is related to the anticyclonic anomaly associated with the SWT as shown in Figure 1. As a result, the strong ascending motion occurs over the tropical eastern Pacific (Figure 3b), which induces strong low‐level westerly winds here (Figure 3c). The strong westerly winds contribute to the development of the ENSO‐like SST by transporting the warmer water from the West Pacific warm pool to the tropical eastern Pacific (e.g., Chen et al., 2016; Huang et al., 2001). In addition, the northerly wind associated with the anticyclonic anomaly in the Northwest Pacific also favors the development of such a westerly wind that warms that SST in the eastern tropical Pacific (Figure 3c). To quantify the contribution of the SWT to ENSO‐like SST anomalies, we compute the SST anomalies associated with PC2 (representing the SWT) and those regressed onto the Niño3.4 index (representing the internal ENSO variability), and compare their amplitudes in the eastern tropical Pacific (Figure S5 in Supporting Information S1). It is found that the SST anomaly attributed to the SWT accounts for 34%–48% of ENSO‐related SST anomaly over the eastern tropical Pacific (Figure S5 in Supporting Information S1). These results suggest that the SWT contributes to the amplitude of the ENSO‐related SST. The abovementioned results suggest that the SWT is indeed crucial for the development of the ENSO‐like SST from both statistical and physical mechanism. Moreover, from Figures 1 and 3a, we find that the SWT may

📷

Figure 3. (a) Regression maps of the divergence wind (vectors; unit: m s− 1) and velocity potential (shading; unit: 106 m s− 1) at 250 hPa during NDJ of 1979–2019 against the PC2. (b) As in (a), but for the height‐longitude cross‐section of vertical wind field (zonal wind and omega; vectors) and omega (shading; unit: Pa s− 1) along the equator. (c) As in (a), but for 850‐hPa wind field (black vectors; unit: m s− 1) and sea surface temperature (SST) (shading; unit: (k). The 850‐hPa winds are marked magenta when significant at a 90% confidence level using the Student's t test. White dots indicate that the omega (b) and SST (c) signals are significant at the 90% confidence level using the Student's t‐test. A, B, C, …, J denote the central positions of the wave train as shown in Figure 1. originate in the North Atlantic. This suggests that the SWT may be associated with variability in the North Atlantic. The Formation Mechanism of the Wave Train Along the Subtropical Westerly Jet In this section ,we examine the drivin gmech an ismof the SWTan dit spossible connection with theSSTanomalyin the northeastern Atlantic. Accordingto previous studies(i.e., An,Sheng, et al., 2022; Luoet al., 2022; Watanabe & Kimoto, 2000), diabatic heating (refers to q1 in this study) released by rainfall may force an atmospheric response. Therefore, we show distributions of precipitation related to the SWT (Figure 4a). The positive precipitation anomaly is clearly evident over southwestern Europe and its coast when the SWT appears (Figure 4a). Meanwhile, a cyclonic anomaly appears in the North Atlantic (Figure 4a). According to Shaman and Tziperman (2011), these

📷

Figure 4.

precipitation anomalies over southwestern Europe are associated with the cyclonic anomaly in the North Atlantic that alters low‐level westerly winds and onshore moisture advection from the Atlantic. According to the thermal wind principle, the southerly wind anomalies associated with the cyclonic anomaly are likely driven by the SST gradient between cold and warm SST anomalies in the North Atlantic (Figure3c).Inaddition ,according to Shaman and Tziperman (2011), ENSO may modulate such precipitation and the related atmospheric circulation. In this study, however, the ENSO signal does not reach the North Atlantic (Figure 1, S2). Therefore, we speculate that the local SST anomaly in the North Atlantic could be crucial (Figure 3c).

We thus further examine the causality between the SST anomaly and the precipitation anomaly according to the L‐K information flow (Liang, 2014; Text S4). The precipitation index used is the area‐averaged precipitation anomaly in the domain (35–55°N, 20°W–0°). The SST index is defined as the difference of the SST anomaly between (30− 60°N, 40°W− 0°) marked by blue box and (30− 60°N, 0°− 40°E) marked by red box in Figure 4b. These two regions correspond well to the dipole‐like SST anomaly pattern observed over the North Atlantic in Figure 3c, which suggests that this index effectively captures the key SST variability associated with the SWT pattern. Applying Equation 4 in Text S4 to the anomalous precipitation and SST fields gives a flow of information from the precipitation to the SST anomaly: TP→S = − 7.9 (unit: 10− 3 nats year− 1), and a flow from the SST to precipitation anomaly: TS→P = − 48 (unit: 10− 3 nats year− 1). In other words, the precipitation and the SST anomaly are mutually causal, and the causality is asymmetric, with the one from the latter to the former larger than its counterpart. The negative sign of TS→P implies that the SST gradient over the northeastern Atlantic tends to promote rainfall here. Moreover, we find that the cyclonic anomaly and the associated precipitation anomaly over the northeastern Atlantic are indeed related to the SST gradient in the northeastern Atlantic and the coast of Europe based on a linear regression analysis (Figures 3c and 4b) and CAM6.3 simulations (Text S5, Figures S6 and S7 in Supporting Information S1), which further verifies our hypothesis from the physical mechanism. That is to say, the SST gradient may strengthen the cyclonic anomaly by inducing anomalous westerly and southerly wind along the coast of Western Europe (Figure 3c), which eventually leads to a positive precipitation anomaly (Figures 4a and 4b), and vice versa. Interestingly, this strong connection between the precipitation and the SST gradient mainly occurs during November to the following February (Figure S8 in Supporting Information S1), which may be one of the causes of the seasonal ENSO phase locking. The question now is whether the diabatic heating associated with rainfall over the northeastern Atlantic can trigger a SWT‐like pattern.

To distill the key physical mechanisms for the formation of the SWT, we calculated q1 according to Yanai et al. (1973). As shown in Figures 4c and 4d, a large amount of diabatic heating is released by the rainfall process, which presents a gamma distribution with height. Based on such an observational heating source, we conducted numerical experiments using the LBM. In the experiment, the center of heating forcing is located at 700 hPa at 10°W, 45°N, with a horizontal extent spanning from 30°W to 10°E in the longitudinal direction and from 35 to 55°N in the latitudinal direction as shown in Figure 4e. An evident wave train emanates from West Europe and propagates eastward to eastern Pacific along the subtropical westerly jet in LBM (Figure 4e‒4f), which is consistent with that in observations (Figure 1). In particular, the LBM used replicates the anticyclonic anomaly of the eastern Pacific in observations well, which is responsible for the ENSO‐like SST (Figure 4e‒4f). Interestingly, the results based on CAM6.3 also supports these results from observations and the LBM (Figure S9 in Supporting Information S1). Notably, we highlight in this study the triggering role of the northeastern Atlantic SST anomaly on the SWT. According to An, Chen, et al. (2022), the internal variability also plays an important role in maintaining this wave train through barotropic and baroclinic energy conversions due to it propagating along the subtropical westerly jet.

In summary, our results suggest that the SWT favors the development of the ENSO‐like SST in boreal winter, while the former is triggered by the diabatic heating released by rainfall. These precipitation anomalies may be related to the SST gradient in the northeastern Atlantic.

Figure 4. (a) Regressed precipitation (shading; unit: mm day− 1) and 850‐hPa wind (vectors; unit: m s− 1) during NDJ of 1979–2019 onto the PC2. (b) Same as (a), but for onto the standardized sea surface temperature (SST) index (SST difference between the blue and red box in Panel (4b). (c) Regressed 500‐hPa q1 (shading; K day− 1) onto the standardized precipitation index (shading; unit: m s− 1) over 35–55°N, 20°W–0°. (d) Vertical profile for the area‐averaged q1. White dots indicate that the precipitation (a–b) and q1 (c) signals are significant at the 90% confidence level using the Student's t‐test. (e) Time‐mean stream function at day 1–10 in LBM (shading; unit: 106 s− 1) and regressed stream function (black contours; unit: 106 s− 1) during NDJ of 1979–2019 against the PC2. (f) As in (e), but for day 1–20. The magenta contours represent the diabatic heating at 500 hPa in the LBM. The day 1–20 period was chosen to better illustrate the full wave train pattern in the figure.

In this study, we examine a new possible link between the North Atlantic and the ENSO‐like SST through the SWT in boreal winter. The results show that a positive (negative) SST gradient (difference between (30− 60°N, 40°W− 0°) and (30− 60°N, 0°− 40°E)) in the northeast Atlantic is conducive to a positive (negative) precipitation anomaly here by strengthening a cyclonic anomaly here, which transports much moisture for rainfall. The L‐K information flow confirmed this cause‐effect relation. The diabatic heating released by rainfall serves as a heating (cooling) forcing to excite an anticyclonic anomaly in the upper troposphere there, which propagates eastward in the form of the Rossby wave along subtropical westerly jet (i.e., the SWT). It leads to an anticyclonic anomaly in the upper troposphere of the eastern Pacific. This anticyclonic anomaly causes a strong divergence (convergence) there and thus promotes (inhibits) atmospheric convection. The promoted (inhibited) convection induces anomalous westerly (easterly) winds over the equator, which contributes to the development of El Niño (La Niña)‐like SST. The SWT formation process is also reproduced by the LBM and CAM6.3. The schematic diagram of the physical mechanism is shown in Figure S10 in Supporting Information S1.

Compared with previous studies (Barnston & Livezey, 1987; C. Z. Wang, 2019; X. Wang et al., 2013), this study is the first attempt to investigate the linkage between the North Atlantic and ENSO through the wave train along the subtropical westerly jet, which may be of great importance for understanding the development of ENSO and its phase locking. Notably, all three super El Niño events (1982, 1997, and 2015) are in the second quadrant of Figure 2, suggesting that the SWT may play an important role in the development of the super El Niño. J. Wang & Wang (2021) highlighted the super El Niño as an outcome of pantropical interactions. However, this study provides a new perspective to consider the super El Niño from extratropical interactions. In addition, the relationship between the North Atlantic and ENSO via the SWT is synchronous in the present study. According to previous studies (e.g., An, Chen, et al., 2022, An et al., 2020; Huang et al., 2020; Li & Sun, 2015; Luo et al., 2022), the SWT‐like pattern is one of the common patterns in the Northern Hemisphere. Hence, it would be of interest to predict ENSO events if the lag‐lead relation between the North Atlantic and ENSO via the SWT is found. Moreover, as shown in Figures 1 and 3, the SWT may also modulate the development of the Victoria mode (Ji et al., 2024) and warm blobs (Hu et al., 2025) over the North Pacific, which deserves more studies in the future.

It is worth noting that a limitation of this study is the lack of further investigation into the causes of the North Atlantic SST gradient. Our information flow results indicate precipitation receives information flow from the SST gradient, while the SST gradient also receives information flow from precipitation (i.e., atmospheric signals). This bidirectional exchange suggests a complex air–sea coupling in the key region identified. This warrants further investigation in future studies. Moreover, as mentioned in the Introduction, the wave train focused on in this study corresponds to the second EOF mode of 200 hPa meridional wind anomalies (unit: m s− 1) over domain (80°W− 160°E, 5− 80°N) during NDJ (i.e., November− December 1979 and January 1980) from 1979 to 2019. The first mode is primarily driven by ENSO and serves as an important bridge for transmitting ENSO signals to the North Atlantic (An, Chen, et al., 2022). This mode has been thoroughly discussed in the study by An, Chen, et al. (2022). For more details, please refer to An, Chen, et al. (2022).