Did removing the ENSO effect from global sea surface temperature (SST) variability cause the unprecedented marine heat waves of 2023-2024? Why has global climate change and global warming increased?

The El Niño–Southern Oscillation (ENSO) is a dominant driver of seasonal–interannual climate variability and has been associated with unprecedented extreme conditions such as marine heat waves (MHWs). However, quantifying the effects of ENSO on MHW characteristics remains a challenge due to data limitations. Here, we use a set of “pulse-forming” simulations of the tropical Pacific Ocean with a fully coupled Earth system model as a platform to assess the skill of four empirical methods aimed at isolating the contribution of ENSO to monthly SST anomalies including MHW extreme conditions. We then applied the most skilled method to the observational record to determine the impact of ENSO on the spatial coverage, intensity, and duration of MHWs since 1960 (after removing the background warming trend). We find that the 2023–24 El Niño contributed to about half of the global coverage of the record-breaking MHWs, with the tropical Indian and tropical Atlantic Oceans having the greatest impact. Our results highlight the critical role of ENSO in generating the most extreme MHW conditions on record.

Statement of Significance: In 2023 and early 2024, global ocean surface temperatures reached unprecedented highs, leading to marine heat waves that spanned large areas of the global ocean.

These waves coincided with an El Niño event, prompting us to investigate the extent to which El Niño events influence global marine heat waves. To isolate the contribution of El Niño to these heat waves, we compared the effectiveness of several methods in removing the influence of El Niño from ocean surface temperature data. Using the best of these methods, we found that the 2023-24 El Niño was responsible for about half of the spatial coverage of simultaneous marine heat waves, with the remainder due to unrelated climate variability.

1.Introduction:

The El Niño–Southern Oscillation (ENSO) phenomenon, through its extensive teleconnections, influences sea surface temperature (SST) variability beyond the tropical Pacific Ocean, making it one of the most important drivers of extreme water heat waves or marine heat waves (MHWs) (Oliver et al., 2018). Extreme heat waves can have severe ecological and socioeconomic impacts; hence, studying their drivers is of critical importance for climate resilience and adaptation. The relationship between ENSO and extreme heat waves received particular attention in 2023 and early 2024, as record-breaking ocean temperatures were observed worldwide in conjunction with a strong El Niño event (Huang et al., 2024; Jiang et al., 2024, 2025; Johnson et al., 2024), leading to the establishment of a causal relationship between these phenomena. One way to quantify this possible association is to statistically remove the ENSO signal from the sea surface temperature anomalies (SSTA) in 2023–24 and then recalculate the MHW features. Many approaches to removing the ENSO signal from climate data have been proposed over the years, for example, through linear regression on the ENSO index (Chiang and Vimont 2004; Robock and Mao 1995; Santer et al. 2001), removing empirical special cases related to ENSO (Compo and Sardeshmukh 2010; Huang et al. 2024; Kelly and Jones 1996), building a stochastic climate model for SST trends that explicitly resolves the ENSO forcing (Gunnarson et al. 2024), and using a linear inverse model (LIM) to filter the dynamical evolution of ENSO (Solomon and Newman 2012). However, a systematic comparison of the effectiveness of these methods has not yet been conducted.

In this study, we used a set of simulations of the Tropical Pacific Pacemaker of the Community Earth System Model Version 2 (CESM2) to judge how well each of the four methods above removes the “true” ENSO signal from the simulated SSTAs. We then applied the most skilled of these methods to observational data to estimate the influence of ENSO on record-breaking MHWs since 1960, with a special emphasis on the El Niño event in 2023–24. Since this work focuses on domestic climate variability and its relationship to ENSO, we examined the MHWs defined after removing the anthropogenic warming trend.

2. Data and Methods

A. Tropical Pacific Pacemaker Simulations and Observational Dataset. To evaluate the effectiveness of different ENSO removal methods, we used the CESM2 Tropical Pacific Ocean Pacemaker Experiments (CESM2-TPACE). CESM2-TPACE is based on CESM2 (Danabasoglu et al., 2020) and includes 10 ensemble members from 1880 to 2019 with a horizontal resolution of approximately 1 degree, each with slightly different initial conditions and with historical radiative forcing from 1880 to 2014 and a common socio-economic path of 3–7.0 (SSP3–7.0) from 2015 to 2019. In addition, the tropical Pacific SSTAs of each member were guided to observations using the Extended Reconstructed SST dataset, version 5 (ERSSTv5; Huang et al., 2017), while the rest of the world was freely evolved (an outline of the guidance region is shown in Appendix Figures A1 and A2). As a result, outside the excitation region, the SSTA variability in the CESM2-TPACE ensemble mean ⟨T'⟩ can be considered the true impact of the observed SSTA variability in the tropical Pacific on global climate (as simulated by the model) plus the impact of the common external radiative forcing for each member (see also Deser et al., 2017). To isolate the ENSO contribution, we remove the radiative forcing component by subtracting the mean of the CESM2 large ensemble (CESM2-LE; Rogers et al., 2021) at each location and time, using 50 members driven by the same radiative forcing as used in CESM2-TPACE (i.e., members that used the CMIP6 biomass burning protocol). The time series of “ENSO-free” SSTAs in each CESM2-TPACE member at each location and time can be calculated as Tef= -T -T ,

where T' is the total SSTA and ¢TEF represents the simulated SSTA without ENSO, independent of the influence of ENSO and external radiative forcing according to the procedures outlined above. The ¢TEF variable, in particular, is a useful basis for comparing any ENSO removal technique. Specifically, we apply each ENSO removal method to T' and generate several estimates of ¢TEF that can be used to quantitatively assess the effectiveness of each method. However, ⟨T'⟩ contains a non-zero residual of non-ENSO internal variability due to the small ensemble size (Rowell et al., 1995), which we corrected for (see Appendix for more information). Figure 1 shows an example of T' and ⟨T'⟩ for a single location in CESM2-TPACE.

We emphasize that the pacemaker experiment was used only to test the different methods. CESM2 shows a realistic ENSO even in the presence of (and) remote coupling.

Figure 1. SSTA time series for ENSO and the Indian Ocean in the CESM2-TPACE experiment. (a) Mean SST Nino-3.4 index (black curve; note that all members of the ensemble have approximately the same ENSO indices). Red and blue shadings indicate El Nino and La Niña events, respectively (events are defined as when the Nino-3.4 index is one standard deviation greater or less than minus one standard deviation, respectively). (b) SSTA time series at 30°S, 80°E in the Indian Ocean, chosen as an example of a location with a high correlation with the Nino-3.4 index. Individual ensemble members (T') are shown in gray, and the ensemble mean (⟨T'⟩) is shown in black. The red curve indicates the seasonal 90th percentile threshold used to calculate the MHWs.(somewhat biased) ENSO dynamics (Capotondi et al., 2020); hence, we consider CESM2-TPACE as a realistic testbed for our methods. For the results presented in Section 3, these methods were applied to observational data and therefore do not transfer any model biases present in CESM2-TPACE to them.

In assessing the observed MHW conditions, we used the Hadley Centre Sea Ice and Sea Surface Temperature v1.1 (HadISST) dataset of monthly mean SST with a spatial resolution of 1° (Rayner et al., 2003) from January 1960 to November 2024. Monthly anomalies were calculated by removing the climatology for each month separately and then removing the background trend by subtracting the lowest eigenmode of the attenuation using LIM analysis (see Xu et al., 2022).

We did not analyze grid points that had missing SST data (e.g., from sea ice cover). We also used sea surface pressure data from the fifth generation reanalysis produced by ECMWF (ERA5; Hersbach et al., 2020), for which anomalies were calculated by subtracting the monthly climatology and a linear trend.

B. Methods for removing the effect of ENSO. A brief summary of each of the four methods is provided below; details can be found in the Appendix.

1) Linear regression on the SST index Nino-3.4. The simplest (and most widely used) method for removing ENSO is to assume that SSTAs consist of a component due to internal variability independent of ENSO and a component that is a linear function of ENSO status, in this case represented by the SST index of Niño-3.4 (mean SSTAs at 5°N-5°S, 170°-120°W; Barnston et al., 1997). The ENSO component of SSTAs is found through linear regression on the SST index of Niño-3.4. Due to the large thermal inertia of the ocean mixed layer, the maximum correlation between SSTAs and Niño-3.4 usually occurs when SSTAs lag Niño-3.4 by several months (e.g., Alexander et al., 2002). Therefore, we evaluated the regression method at two different lags: 0 and 3 months.

2) Removing eoF. The second method is a more general version of the first method, which uses the principal component (PC) time series associated with the spatial patterns of the empirical orthogonal function (EOF) leading global SSTAs instead of the SST index of Niño-3.4 units (Huang et al., 2024; Kelly and Jones, 1996). We evaluated the optimal number of EOFs/PCs to remove based on the criteria outlined in Section 2c. 3) Trend regression. The third method uses an extension of the original stochastic climate model developed by Hasselman (1976):

where l is the feedback coefficient (attenuation), b is the ENSO teleconnection coefficient,

N(t) is the ENSO index (in this case Niño-3.4), and ξ(t) is the random forcing (white noise). Similar models have been successfully used to assess the remote influence of ENSO on SST variability modes such as the Pacific decadal oscillation (Newman et al., 2003, 2016; Schneider and Cornuelle, 2005) and the Indian Ocean Dipole (Stuecker et al., 2017; Zhao et al., 2019), as well as on North Pacific SST variability in general (e.g., Gunnarson et al., 2024; Park et al., 2006). Both l and b have seasonally adjusted values, allowing equation (2) to represent dynamics such as ENSO-combined modes (Stuecker et al., 2013). To construct SST anomalies without the influence of ENSO, equation (2) is fitted to the data at each grid point via multiple linear regression and then integrated forward in time without the ENSO teleconnection term.4) LIM filter. The fourth method uses a LIM to construct an optimal perturbation filter - following Solomon and Newman (2012). LIM assumes that the SST dynamics can be represented as a linear system driven by a random force (Penland and Sardeshmukh 1995):

where x is the state vector of the system (i.e., the SSTAs at different times and locations in this study), L is the dynamical operator matrix describing the dynamical properties of the evolution of x, and ξ is the random force vector. For an initial state x(t), the most likely state.

The LIM method objectively determines the “optimal initial condition” that translates into a specific final condition (e.g., mature ENSO). The L matrix is generally non-normal, with orthogonal eigenvectors (Penland and Matrosova 2006), which allows for transient anomaly amplification through eigenmode constructive interference before the anomalies eventually dissipate. The optimal initial condition is the one that maximizes the mode vector amplification (Penland and Sredshmuk 1995). Following Solomon and Newman (2012), we use LIM to remove the ENSO signal from SSTAs by filtering out the variability that evolves from the optimal initial condition of one mature ENSO event to the next.

C. Evaluation Criteria. We used two metrics to evaluate the performance of each method in removing the influence of ENSO from the SSTA field: 1) lagged correlation with the Niño-3.4 index with ENSO-removed SSTAs, and 2) comparison of the variance of the ENSO-removed SSTA field with the SSTAs without ENSO from the pacemaker experiment [see equation (1)]. A perfect method would show zero correlation with Niño-3.4 at all lags and would have the same variance as the SSTAs without ENSO.

D. MHW Definitions. MHWs can be defined in several ways, particularly in the case of fixed versus variable baselines (Amaya et al., 2023; Smith et al., 2025). Since our study is concerned with the influence of ENSO on domestic climate variability and not with mean-mode change, we use a variable baseline (i.e., by subtracting the pattern of mean-mode change; see section 2a). In this study, we defined MHW as any month that exceeds the 90th percentile of SSTAs for that calendar month in a given grid cell. We calculated MHW thresholds using complete SSTAs (i.e., before removing the effect of ENSO). MHW duration is defined as the number of consecutive months in a grid cell that experience MHW conditions (note that this definition is different from the definition of daily data). Due to the limited observation time frame, we spatially smoothed the data by calculating the mean and maximum (97.5th percentile) of MHW durations in 5° × 5° bins. MHW intensity

was calculated by summing the SSTAs of MHWs in a grid cell and then dividing by the total number of months in the dataset (expressed in °C-weeks per year -1, analogous to degree-warming weeks). Since this is a unified value that is less noisy than the duration statistics, spatial disaggregation is unnecessary.

3. Results

A. ENSO fingerprint. Of the four methods for removing the influence of ENSO on SSTAs mentioned in Section 2b, SST trend regression and LIM filter performed best and showed comparable skill (see discussion in Appendix). Because the trend regression method Since it is much simpler to implement and performs slightly better at removing the ENSO effect than the LIM filter method, we present the results based on that approach in the main text; however, very similar results were found using the LIM filter (Figs. S1–S3 in the online supplementary material). The linear regression method (with a 3-month pretest) reasonably removed the Niño-3.4 correlation, but its variance was greater than the true ENSO-free variance. The EOF re-mobile method (with two or three EOFs removed) reproduced the ENSO-free variance best of all the methods, but performed worse than the trend regression or LIM filter methods when removing the Niño-3.4 correlation. Based on these results, we applied the trend regression method to the observed SSTAs to quantify the influence of ENSO on MHWs over the historical record (since 1960) as well as the influence of the most recent El Niño on the record-breaking MHWs observed in 2023 and early 2024.

Although the trend regression method is an empirical statistical approach to quantify the teleconnection influence of ENSO on SSTAs, its results are in good agreement with previous literature on the dynamics of ENSO teleconnections. In particular, the spatial pattern of ENSO-induced SSTAs obtained with the trend method (Fig. 2) is similar to that found in previous studies (e.g., Alexander et al. 2002; Lau and Nath 1994, 1996) with (during El Niño) a “horseshoe” pattern of cooling in the North and South Pacific, warming in the Indian Ocean, and warming in the subtropical North and South Atlantic Oceans with poleward cooling. ENSO-related atmospheric circulation anomalies overlap these warming and cooling patterns: the so-called atmospheric bridge connects extratropical SSTs to the equatorial Pacific through atmospheric Rossby waves. For example, in the North Pacific, a bipolar pattern with cooling in the basin center and warming along the west coast of North America is associated with a deepening of the Aleutian Low. As previous studies have shown, cold, dry air moving along the western side of this atmospheric circulation anomaly cools the ocean during El Niño, and warm, moist air moving along the eastern side of the atmospheric circulation anomaly leads to ocean warming (Alexander et al., 2002). Similar long-distance coupling patterns exist in the South Pacific and the North and South Atlantic Oceans. In the tropical Indian and Atlantic Oceans, Walker circulation modulations link SSTAs to ENSO [see Taschetto et al. (2020) for a review of ENSO-atmosphere teleconnections].

B. ENSO impact on observed MHW duration and intensity. The ENSO impact on observed MHW duration and intensity over the period 1960–2024 is shown in Figure 3.

ENSO increases MHW duration and intensity almost everywhere, consistent with previous work (Oliver et al., 2018). The largest increases in duration and intensity due to ENSO, except for the tropical Pacific itself, occur in regions that are strongly influenced by ENSO.

Figure 2. ENSO teleconnection strength in the HadISST observational dataset using the trend regression method. Teleconnection strength (shaded; K months -1) is the annual mean teleconnection coefficient β times the standard deviation of the El Niño-3.4 index, which represents a typical value for teleconnection forcing during an El Niño event. As discussed in the Appendix, the trend regression method cannot reproduce ENSO itself. Therefore, the teleconnection strength in the tropical Pacific is not physically meaningful.

The contours show the regression of ERA5 sea surface pressure on the El Niño-3.4 index, with dashed contours indicating negative values, a contour interval of 30 Pa K -1 and a thicker line at 0 Pa K -1

.Figure 3. ENSO influence on the mean and maximum duration (97.5th percentile) of MHWs and the mean intensity of MHWs during 1960–2024 based on the HadISST dataset. (Top) observed, (middle) after removing the ENSO influence using the trend regression method, and (bottom) their difference (i.e., the ENSO influence, shown as a percentage of the removed ENSO value). Duration is calculated using 5° × 5° bins. MHW intensity is the sum of MHW SSTAs divided by the number of years in the data in a given grid cell (°C-weeks yr−1). Note the different ranges of the color bar in the difference panels.Teleconnections, especially in the tropical parts of the Indo-Atlantic Ocean, and a horseshoe pattern in the northern and southern Pacific Ocean. On average in the world's oceans between 60°S and 60°N (excluding the tropical Pacific: 20°S to 15°N, 160°E to 70°W), ENSO increases the mean duration of MHW by 9.2% and the duration of extreme MHW (97.5th percentile) by 23.9%. In regions with strong ENSO influence, the increase in mean MHW duration can be more than 50% and the duration of the 97.5th percentile can increase by more than 100%. The latter case illustrates the influence of ENSO on very persistent but rare MHWs (see Fig. S4). The global mean MHW intensity without ENSO is 2.27°C/week/year and with ENSO is 2.78°C/week/year, representing a 22.5% increase, although in regions with high teleconnection strength in the Pacific and Indian Oceans, the increase in intensity can exceed 100%.

C. ENSO effect on observed MHW spatial coverage. The area of global ocean covered by MHWs at any given time is also strongly affected by ENSO. By defining MHWs using the 90th percentile threshold, we expect approximately 10% of the ocean to experience MHW at any given time due to chance. However, our results show that El Niño tends to greatly increase the observed MHW spatial coverage beyond the tropical Pacific (Fig. 4). To quantify this effect, we define Aobs and ATR as the percentage of the global ocean area (60°S to 60°N, excluding the equatorial Pacific) with MHW conditions before and after ENSO removal (using the trend regression method), respectively. We also define the “signal-to-noise ratio” asThe signal-to-noise ratio (SNR) represents the contribution of ENSO to the global MHW region relative to the contribution of uncorrelated internal variability.

During El Niño events, Aobs increases with peak values between 21% and 27% (24.6% in February 2024; Fig. 4b, black curve). These increases are much reduced or absent for ATR (Fig. 4b, green curve), as confirmed by the difference Aobs - ATR (Fig. 4c). The maximum.

The MHW regional coverage typically lags the peak of an El Niño event by several months, likely due to the oceanic effect that integrates the atmospheric forcing resulting from the ENSO teleconnection.

La Niña events do not appear to have a significant impact on MHW coverage, likely due to their smaller amplitude compared to El Niño events. In other words, only El Niño generates a signal large enough to clearly emerge from the background noise of the climate system for this particular measure of global MHWs (note that La Niña can induce MHWs on a regional scale; e.g., Feng et al. 2013).

The SNR also peaks during El Niño, although it suggests that some of the increase in global MHW coverage is due to the confluence of ENSO effects and stochastic internal variability (Fig. 4d). For example, the 1987–1988 El Niño had a maximum SNR of 1.0, indicating that the coverage of MHW caused by El Niño was directly comparable in magnitude to that caused by other forms of internal variability. In contrast, the 2015–2016 El Niño had the highest observed SNR of 2.2, so about two-thirds of the total MHW area was directly caused by El Niño.

The 2023–2024 El Niño had a maximum SNR of 1.1 (in December 2023), with an SNR of 1.0 at the peak of MHW coverage in February 2024 and an average of 0.7 between January and May 2024. This suggests that internal variability unrelated to ENSO was almost as important as El Niño in generating the extensive global MHW coverage during that event. However, the MHW spatial coverage histogram (Fig. 4e) shows that the maximum spatial coverage of MHWs in 2023–24 was beyond that caused solely by internal variability. Therefore, the widespread nature of MHWs in 2023–24 could not have occurred without the influence of El Niño.

D. The influence of El Niño 2023–24. Figure 5 shows the spatial structure and evolution of El Niño 2023–24 and its influence on global MHWs (Fig. S5 shows the same information for historical El Niño events). The influence of ENSO (recall Fig. 2) can be captured by the difference between the observed SSTAs and those calculated through our trend regression method. In the Indian Ocean, El Niño warmed the eastern half of the basin, leading to a positive-phase Indian Ocean Dipole (IOD) event in the fall and winter of 2023, which was then transferred to Figure 4. ENSO impact on spatial coverage of MHWs based on the HadISST dataset. (a) El Niño-3.4 index. El Niño events are shaded in red and La Niña events are shaded in blue. (b) Percentage of the global ocean area (60°S–60°N, excluding the tropical Pacific: 20°S–15°N, 160°E–70°W) with MHW conditions before (Aobs; black curve) and after ENSO removal using the trend regression method (ATR; green curve). (c) Aobs - ATR. (d) ENSO SNR. (e) Probability density function of Aobs (gray) and ATR (green). The MHW Aobs and ATR values for February 2024 are shown by the dotted gray and dashed green lines, respectively.Basin-wide warming in early 2024. This sequence and its association with ENSO are consistent with previous studies (Klein et al. 1999; Saji et al. 1999; Stuecker et al. 2017). When the influence of ENSO is removed, the IOD pattern is significantly reduced, and MHWs remain only in the southern Indian Ocean. The role of El Niño in the North Pacific was to suppress MHWs in the central part of the basin (in November and December 2023) and increase them along the west coast of North America, which is a result of the positive phase of the Pacific Decadal Oscillation that is strongly associated with ENSO (e.g., Newman et al. 2016; Schneider and Cornuelle 2005). MHWs increased in the South Pacific, likely due to Pacific-South American atmospheric variability modes that transmit the influence of ENSO to the South Pacific (Mo 2000; Mo and Higgins 1998).

In the Atlantic, a significant portion of the MHW coverage during 2023–24 appears to have originated from internal variability unrelated to ENSO, as evidenced by the significant MHW area remaining after removing the ENSO effect, particularly in the South Atlantic after February 2024 (Fig. 5, middle column). However, El Niño led to MHWs in the tropical Atlantic as well as MHWs in the subtropical North and South Atlantic (Fig. 5, right column). Warming in the tropical and subtropical North Atlantic been linked to El Niño (e.g., Alexander and Scott 2002; Huang 2004) as has warming in the subtropical South Atlantic (Rodrigues et al. 2011, 2015). Thus, 2023–24 MHWs in the Atlantic, which covered a large part of the basin, originated from a confluence of ENSO and unrelated internal climate variability.

Figure 5. Evolution of the 2023-2024 El Niño and associated MHWs. (Left) Observed SSTAs from HadISST with MHWs highlighted and shaded in black. (Center) SSTAs and MHWs without ENSO influence calculated using the trend regression method.

(Right) Difference between observed SSTAs and MHWs and those removed by ENSO. Areas where ENSO influence caused MHWs are shaded in red and shaded in blue.

4. Conclusions:

In this study, we compared several empirical methods for removing the influence of ENSO from global SSTAs using a “full model” approach and then applied the most skillful method to observed MHWs over the period 1960–2024 after removing the background warming trend. Using the CESM2-TPACE experiments as a testbed, we concluded that the trend regression and LIM filter methods are the most skillful methods in removing the influence of ENSO based on two criteria. The trend regression method is somewhat more flexible, simpler to implement, and does not require large-scale spatial data, which may be useful in cases where additional climate modes or physical processes are desired or for situations where data are sparse (e.g., paleoclimate studies).

Using a trend regression method to remove the influence of ENSO from the observed SST data, we find that ENSO acts to increase the average duration of MHW by 9.2% and the intensity of MHW by 22.5%. The spatial coverage of MHWs increases during El Niño events, typically reaching 21–27% of the global ocean outside the tropical Pacific, compared to 10% or less in non-El Niño years. By removing the influence of ENSO, we confirm that these increases are indeed caused by El Niño events. About half of the spatial coverage of MHWs in 2023–2024 was caused by the concurrent El Niño. In fact, we find that the widespread nature of the MHWs in February 2024 could not have occurred without the influence of El Niño.

The evolving spatial pattern of MHWs during the 2023–2024 El Niño closely follows previously studied global ENSO teleconnections. During 2023–24, Indian Ocean heat waves (MHWs) were found to occur over much of the Indian Ocean due to El Niño influence, as were the suppression and enhancement of North Pacific heat waves and the enhancement of South Pacific heat waves. Atlantic heat waves, which covered a large part of the basin, were driven by the confluence of ENSO teleconnections and unrelated internal variability. Our results highlight the importance of ENSO in driving marine heat waves around the world. However, there is little agreement among climate models on how ENSO responds to anthropogenic forcings and how it responds (Maher et al. 2023). Therefore, accurate modeling of ENSO dynamics is crucial to understanding how Indian Ocean heat waves will change as the world warms.

References:

Alexander, M., and J. Scott, 2002: The influence of ENSO on air-sea interaction

in the Atlantic. Geophys. Res. Lett., 29, 1701, https://doi.org/10.1029/2001GL

014347.

Alexander, M. A., and C. Deser, 1995: A mechanism for the recurrence of winter-

time midlatitude SST anomalies. J. Phys. Oceanogr., 25, 122–137, https://doi.

org/10.1175/1520-0485(1995)0252.0.CO;2.

——, I. Bladé, M. Newman, J. R. Lanzante, N.-C. Lau, and J. D. Scott, 2002: The

atmospheric bridge: The influence of ENSO teleconnections on air–sea inter-

action over the global oceans. J. Climate, 15, 2205–2231, https://doi.org/10.

1175/1520-0442(2002)0152.0.CO;2.

Amaya, D. J., and Coauthors, 2023: Marine heatwaves need clear definitions so

coastal communities can adapt. Nature, 616, 29–32, https://doi.org/10.1038/

d41586-023-00924-2.

Barnston, A. G., M. Chelliah, and S. B. Goldenberg, 1997: Documentation of a

highly ENSO-related SST region in the equatorial Pacific. Atmos.–Ocean, 35,

367–383, https://doi.org/10.1080/07055900.1997.9649597.

Capotondi, A., C. Deser, A. S. Phillips, Y. Okumura, and S. M. Larson, 2020: ENSO

and Pacific decadal variability in the community Earth system model ver-

sion 2. J. Adv. Model. Earth Syst., 12, e2019MS002022, https://doi.org/10.

1029/2019MS002022.

Chiang, J. C. H., and D. J. Vimont, 2004: Analogous Pacific and Atlantic meridional

modes of tropical atmosphere–ocean variability. J. Climate, 17, 4143–4158,

https://doi.org/10.1175/JCLI4953.1.

Compo, G. P., and P. D. Sardeshmukh, 2010: Removing ENSO-related variations

from the climate record. J. Climate, 23, 1957–1978, https://doi.org/10.1175/

2009JCLI2735.1.

Danabasoglu, G., and Coauthors, 2020: The Community Earth System Model ver-

sion 2 (CESM2). J. Adv. Model. Earth Syst., 12, e2019MS001916, https://doi.

org/10.1029/2019MS001916.

De Elvira, A. R., and P. Lemke, 1982: A Langevin equation for stochastic climate

models with periodic feedback and forcing variance. Tellus, 34, 313–320,

https://doi.org/10.1111/j.2153-3490.1982.tb01821.x.

Deser, C., I. R. Simpson, K. A. McKinnon, and A. S. Phillips, 2017: The Northern

Hemisphere extratropical atmospheric circulation response to ENSO: How

well do we know it and how do we evaluate models accordingly? J. Climate,

30, 5059–5082, https://doi.org/10.1175/JCLI-D-16-0844.1.

Ebisuzaki, W., 1997: A method to estimate the statistical significance of a correla-

tion when the data are serially correlated. J. Climate, 10, 2147–2153, https://

doi.org/10.1175/1520-0442(1997)0102.0.CO;2.

Feng, M., M. J. McPhaden, S.-P. Xie, and J. Hafner. La Niña forces unprecedented

Leeuwin Current warming in 2011. Sci. Rep., 3, 1277, https://doi.org/10.1038/

srep01277.

Frankignoul, C., 1985: Sea surface temperature anomalies, planetary waves, and

air-sea feedback in the middle latitudes. Rev. Geophys., 23, 357–390, https://

doi.org/10.1029/RG023i004p00357.

Gunnarson, J. L., M. F. Stuecker, and S. Zhao, 2024: Drivers of future extratropical

sea surface temperature variability changes in the North Pacific. npj Climate

Atmos. Sci., 7, 164, https://doi.org/10.1038/s41612-024-00702-5.

Hasselmann, K., 1976: Stochastic climate models Part I. Theory. Tellus, 28,

473–485, https://doi.org/10.1111/j.2153-3490.1976.tb00696.x.

Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy.

Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.

Huang, B., 2004: Remotely forced variability in the tropical Atlantic Ocean.

Climate Dyn., 23, 133–152, https://doi.org/10.1007/s00382-004-0443-8.

——, 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.

——, X. Yin, J. A. Carton, L. Chen, G. Graham, P. Hogan, T. Smith, and H.-M. Zhang,

2024: Record high sea surface temperatures in 2023. Geophys. Res. Lett., 51,

e2024GL108369, https://doi.org/10.1029/2024GL108369.

Jiang, N., and Coauthors, 2024: Enhanced risk of record-breaking regional

temperatures during the 2023–24 El Niño. Sci. Rep., 14, 2521, https://doi.

org/10.1038/s41598-024-52846-2.

——, C. Zhu, Z.-Z. Hu, M. J. McPhaden, T. Lian, C. Zhou, W. Qian, and D. Chen,

2025: El Niño and sea surface temperature pattern effects lead to histori-

cally high global mean surface temperatures in 2023. Geophys. Res. Lett., 52,

e2024GL113733, https://doi.org/10.1029/2024GL113733.

Johnson, G. C., and Coauthors, 2024: Global oceans. Bull. Amer. Meteor. Soc., 105

(8), S156–S213, https://doi.org/10.1175/BAMS-D-24-0100.1.

Kelly, P. M., and P. D. Jones, 1996: Removal of the El Niño-Southern Oscillation

signal from the gridded surface air temperature data set. J. Geophys. Res.,

101, 19 013–19 022, https://doi.org/10.1029/96JD01173.

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. Cli-

mate, 12, 917–932, https://doi.org/10.1175/1520-0442(1999)0122.0.CO;2.

Larson, S. M., D. J. Vimont, A. C. Clement, and B. P. Kirtman, 2018: How mo-

mentum coupling affects SST variance and large-scale Pacific climate

variability in CESM.J.Climate,31,2927–2944,https://doi.org/10.1175/JCLI-D-17-

0645.1.

Lau, N.-C., and M. J. Nath, 1994: A modeling study of the relative roles of

tropical and extratropical SST anomalies in the variability of the global

atmosphere-ocean system. J. Climate, 7, 1184–1207, https://doi.org/10.

1175/1520-0442(1994)0072.0.CO;2.

——, and ——, 1996: The role of the “atmospheric bridge” in linking tropical

Pacific ENSO events to extratropical SST anomalies. J. Climate, 9, 2036–2057,

https://doi.org/10.1175/1520-0442(1996)0092.0.CO;2.

Leith, C. E., 1973: The standard error of time-average estimates of climatic means.

J. Appl. Meteor., 12, 1066–1069, https://doi.org/10.1175/1520-0450(1973)

0122.0.CO;2.

Magorian, T., and C. Penland, 1993. Prediction of Niño 3 sea surface temperatures

using linear inverse modeling. J. Climate, 6, 1067–1076, https://doi.org/10.

1175/1520-0442(1993)0062.0.CO;2.

Maher, N., and Coauthors, 2023: The future of the El Niño–Southern Oscillation:

Using large ensembles to illuminate time-varying responses and inter-model

differences. Earth Syst. Dyn., 14, 413–431, https://doi.org/10.5194/esd-14-

413-2023.

Matsumura, S., G. Huang, S.-P. Xie, and K. Yamazaki, 2010: SST-forced and internal

variability of the atmosphere in an ensemble GCM simulation. J. Meteor. Soc.

Japan, 88, 43–62, https://doi.org/10.2151/jmsj.2010-104.

Mo, K. C., 2000: Relationships between low-frequency variability in the Southern

Hemisphere and sea surface temperature anomalies. J. Climate, 13, 3599–3610,

https://doi.org/10.1175/1520-0442(2000)0132.0.CO;2.

——, and R. W. Higgins, 1998: The Pacific–South American modes and tropical

convection during the Southern Hemisphere winter. Mon. Wea. Rev., 126,

1581–1596, https://doi.org/10.1175/1520-0493(1998)126

2.0.CO;2.

Newman M., G. P. Compo, and M. A. Alexander, 2003: ENSO-forced variability of

the Pacific decadal oscillation. J. Climate, 16, 3853–3857, https://doi.org/

10.1175/1520-0442(2003)0162.0.CO;2.

——, M. A. Alexander, and J. D. Scott, 2011a: An empirical model of tropi-

cal ocean dynamics. Clim. Dyn., 37, 1823–1841, https://doi.org/10.1007/

s00382-011-1034-0.

——, S.-I. Shin, and M. A. Alexander, 2011b: Natural variation in ENSO flavors,

Geophys. Res. Lett., 38, 0094–8276, https://doi.org/10.1029/2011GL047658.

——, and Coauthors, 2016: The Pacific decadal oscillation, revisited. J. Climate,

29, 4399–4427, https://doi.org/10.1175/JCLI-D-15-0508.1.

Nicholls, N., 1984: The Southern Oscillation and Indonesian sea surface tem-

perature. Mon. Wea. Rev., 112, 424–432, https://doi.org/10.1175/1520-0493

(1984)1122.0.CO;2.

Oliver, E. C. J., and Coauthors, 2018: Longer and more frequent marine heat-

waves over the past century. Nat. Commun., 9, 1324, https://doi.org/10.1038/

s41467-018-03732-9.

——, J. A. Benthuysen, S. Darmaraki, M. G. Donat, A. J. Hobday, N. J. Holbrook,

R. W. Schlegel, and A. Sen Gupta, 2021: Marine heatwaves. Ann. Rev. Mar.

Sci., 13, 313–342, https://doi.org/10.1146/annurev-marine-032720-095144.

Park, S., M. A. Alexander, and C. Deser, 2006: The impact of cloud radiative feed-

back, remote ENSO forcing, and entrainment on the persistence of North Pa-

cific sea surface temperature anomalies. J. Climate, 19, 6243–6261, https://

doi.org/10.1175/JCLI3957.1.

Penland, C., and P. D. Sardeshmukh, 1995: The optimal growth of tropical sea

surface temperature anomalies. J. Climate, 8, 1999–2024, https://doi.org/

10.1175/1520-0442(1995)0082.0.CO;2.

——, and L. Matrosova, 2006: Studies of El Niño and interdecadal variability in

tropical sea surface temperatures using a nonnormal filter. J. Climate, 19,

5796–5815, https://doi.org/10.1175/JCLI3951.1.

Preisendorfer, R. W., F. W. Zwiers, and T. P. Barnett, 1981: Foundations of

Principle Component Selection Rules. Scripps Institution of Oceanogra-

phy, 192 pp.

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.

Robock, A., and J. Mao, 1995: The volcanic signal in surface temperature obser-

vations. J. Climate, 8, 1086–1103, https://doi.org/10.1175/1520-0442(1995)

0082.0.CO;2.

Rodgers, K. B., and Coauthors, 2021: Ubiquity of human-induced changes in

climate variability. Earth Syst. Dyn., 12, 1393–1411, https://doi.org/10.5194/

esd-12-1393-2021.

Rodrigues, R. R., R. J. Haarsma, E. J. D. Campos, and T. Ambrizzi, 2011: The

impacts of inter–El Niño variability on the tropical Atlantic and northeast

Brazil climate. J. Climate, 24, 3402–3422, https://doi.org/10.1175/2011

JCLI3983.1.

——, E. J. D. Campos, and R. Haarsma, 2015: The impact of ENSO on the South

Atlantic subtropical dipole mode. J. Climate, 28, 2691–2705, https://doi.

org/10.1175/JCLI-D-14-00483.1.

Rowell, D. P., C. K. Folland, K. Maskell, and M. N. Ward, 1995: Variability of

summer rainfall over tropical North Africa (1906–92): Observations and

modelling. Quart. J. Roy. Meteor. Soc., 121, 669–704, https://doi.org/10.1002/

qj.49712152311.

Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A di-

pole mode in the tropical Indian Ocean. Nature, 401, 360–363, https://doi.

org/10.1038/43854.

Santer, B. D., and Coauthors, 2001: Accounting for the effects of volcanoes and

ENSO in comparisons of modeled and observed temperature trends. J. Geo-

phys. Res., 106, 28 033–28 059, https://doi.org/10.1029/2000JD000189.

Schneider, N., and B. D. Cornuelle, 2005: The forcing of the Pacific decadal oscilla-

tion. J. Climate, 18, 4355–4373, https://doi.org/10.1175/JCLI3527.1.

Smith, K. E., and Coauthors, 2025: Baseline matters: Challenges and implica-

tions of different marine heatwave baselines. Prog. Oceanogr., 231, 103404,

https://doi.org/10.1016/j.pocean.2024.103404.

Solomon, A., and M. Newman, 2012: Reconciling disparate twentieth-century

Indo-Pacific ocean temperature trends in the instrumental record. Nat. Cli-

mate Change, 2, 691–699, https://doi.org/10.1038/nclimate1591.

Sprintall, J., S. Cravatte, B. Dewitte, Y. Du, and A. S. Gupta, 2020: ENSO oceanic

teleconnections. El Niño Southern Oscillation in a Changing Climate, Wiley,

337–360.

Stuecker, M. F., A. Timmermann, F.-F. Jin, S. McGregor, and H.-L. Ren, 2013: A com-

bination mode of the annual cycle and the El Niño/Southern Oscillation. Nat.

Geosci., 6, 540–544, https://doi.org/10.1038/ngeo1826.

——, ——, F. Jin, Y. Chikamoto, W. Zhang, A. T. Wittenberg, E. Widiasih, and

S. Zhao, 2017: Revisiting ENSO/Indian Ocean Dipole phase relationships.

Geophys. Res. Lett., 44, 2481–2492, https://doi.org/10.1002/2016GL

072308.

Taschetto, A. S., C. C. Ummenhofer, M. F. Stuecker, D. Dommenget, K. Ashok,

R. R. Rodrigues, and Y. Sang-Wook, 2020: ENSO atmospheric teleconnections.

El Niño Southern Oscillation in a Changing Climate, 1st ed. Wiley, 311–335.

Timmermann, A., and Coauthors, 2018: El Niño-Southern Oscillation complexity.

Nature, 559, 535–545, https://doi.org/10.1038/s41586-018-0252-6.

Xu, T., M. Newman, A. Capotondi, S. Stevenson, E. Di Lorenzo, and M. A. Alexan-

der, 2022: An increase in marine heatwaves without significant changes in

surface ocean temperature variability. Nat. Commun., 13, 7396, https://doi.

org/10.1038/s41467-022-34934-x.

Zhao, S., F.-F. Jin, and M. F. Stuecker, 2019: Improved predictability of the Indian

Ocean Dipole using seasonally modulated ENSO forcing forecasts. Geophys.

Res. Lett., 46, 9980–9990, https://doi.org/10.1029/2019GL084196