Impact of mean state on intermodel spread of western North Pacific anomalous anticyclone during El Niño mature winter in CMIP6 models

Jiahao Lu; Xieyuan Wang; Yuhan Gong; Zhiheng Chen; Zongci Huang; Xu Chen

doi:10.1038/s41598-025-04048-7

. 2025 Jun 4;15:19539. doi: 10.1038/s41598-025-04048-7

Impact of mean state on intermodel spread of western North Pacific anomalous anticyclone during El Niño mature winter in CMIP6 models

Jiahao Lu ^1,^✉, Xieyuan Wang ^2,^✉, Yuhan Gong ³, Zhiheng Chen ⁴, Zongci Huang ^5,⁶, Xu Chen ¹

PMCID: PMC12137608 PMID: 40467722

Abstract

The impact of the mean state on simulated western North Pacific anomalous anticyclone (WNPAC) during El Niño mature winter was studied through the diagnosis of 45 historical climate experiments from the CMIP6. The result indicates that in addition to the model-simulated El Niño-related sea surface temperature anomalies (SSTA), the atmospheric mean state also plays an important role in modulating the intermodel spread of WNPAC, especially the intensity and zonal location. On one hand, the central Pacific mean-state precipitation/moisture fields can modulate the western North Pacific (WNP) thermal condition responded to the SST forcing in the Niño-3.4 region through a local circulation-convective feedback, and thus affecting the intensity of WNPAC. On the other hand, the mean-state meridional vorticity gradient in the WNP and the westward-extended central Pacific warming may have acted together to regulate the WNPAC zonal location.

Keywords: Western North Pacific anomalous anticyclone, Intermodel spread, Simulation bias, Mean state

Subject terms: Climate sciences, Atmospheric dynamics

Introduction

The western North Pacific anomalous anticyclone (WNPAC), which is also referred to as the Philippine Sea anomalous anticyclone or anomalous northwest Pacific anticyclone, is an important atmospheric circulation system that conveys El Niño impact on East Asian climate^1–7. It has great impacts on East Asian monsoon and also contributes to the decay of El Niño^2–6,8,9.

The life cycle of the WNPAC is tightly linked with both the phase of El Niño and the annual cycle of the tropical climate. It forms during El Niño–developing fall, fully establishes during El Niño mature winter, maintains during the following spring and summer, and decays after that^4,10–15. There have been various theories proposed to understand the formation and maintenance of the WNPAC. They include 1) Warm pool air–sea interaction theory that extends the original local air–sea interaction of Wang et al. to a self-sustained inter-basin coupled mode across the Indian ocean (IO)–western North Pacific (WNP)^4,16, 2) IO capacitor theory^11,12, 3) Combination mode theory¹⁷, 4) Moist enthalpy advection/Rossby wave modulation theory^14,15, and 5) Central Pacific sea surface temperature anomaly (SSTA) forcing mechanism^16,18.

During El Niño mature winter, the anomalous warm SSTA in the central-east Pacific contributes to the onset and development of the WNPAC through a two-step response. Firstly, the positive heating anomaly in the central Pacific stimulates anomalous cyclonic gyre in the WNP through the Rossby wave response. The northeasterly anomalies on the northwest flank of the cyclonic gyre enhance the mean northeasterly trade and cool local SST through enhanced evaporation. Secondly, the cold SSTA contributes to the development of WNPAC through stimulating an anomalous anticyclone to the west. The WNPAC enhances the northeasterly anomalies to its east, and promotes positive wind-evaporation-SST feedback that maintains the cold SSTA^4,10,19,20. On the other hand, the northeasterly anomalies on the northwest flank of the cyclonic gyre also stimulate WNPAC through inducing negative moist enthalpy advection and then cause negative precipitation in the WNP, which is an independent atmospheric process, namely the moist enthalpy advection mechanism¹⁴.

Most of the current state-of-the-art global atmosphere general circulation models (GCMs) are capable of capturing the WNPAC realistically to some extent, while the intermodel differences are still noticeable. A number of previous studies on the source of WNPAC bias during El Niño mature winter and decaying summer is based on Atmospheric Model Intercomparison Project (AMIP) experiments. These AMIP experiments were forced by the same prescribed SST without the interference of SSTA bias. Therefore, the role of atmospheric background field is highlighted. However, there are biases in the amplitude, spatial pattern and frequency of ENSO in coupled models^21–23. As a response to ENSO, the WNPAC is also affected by these biases, hence the SSTA bias is non-ignorable.

After decades of development, the problem that the positive SSTA during El Niño extends westward excessively remains even in the most advanced coupled models^24–27. Such a bias influences the capability of models to simulate precipitation and circulation anomalies in East Asia-WNP during El Niño mature winter^28,29 and decaying summer^30,31. In CMIP6 models, the phenomenon of excessive westward extension of positive SSTA is widespread. It is the oceanic mean state bias that results in the ENSO-related SSTA bias, which can be understood as a physical process that the excessively strong cold tongue in the eastern equatorial Pacific causes the underestimate of wind-SST feedback and shortwave radiation feedback, and then the models cannot reproduce the asymmetry of ENSO and the characteristics of seasonal phase-locking accurately^22,32–35. Previous studies have mainly attributed WNPAC bias to SSTA bias in models. Tao et al.³⁶ adopted the intermodel EOF method and found that the intermodel difference of WNPAC can be divided into two modes: intensity difference and location difference. The former originates from the difference of SSTA in the tropical Western Pacific, while the latter is affected by the SSTA in the tropical Indian Ocean (TIO), the WNP and the central eastern Pacific. Feng et al.³⁷ evaluated the models participating in the CMIP5 historical climate experiment, and found that more than half of the models could not reproduce the WNPAC during the decaying summer of CP El Niño, because the warm SST during El Niño in these models lasted too long. The warm SST in the central Pacific can persist into summer and stimulate the anomalous cyclone in the central and western Pacific, which is not favorable for the maintenance of WNPAC.

Compared to oceanic mean state and SSTA bias, the role of atmospheric background mean state has received less attention. Spencer and Slingo³⁸ used the observed SST to drive the HadAM3 (Hadley Centre Atmospheric Model version 3) atmospheric general circulation model (AGCM), and found that the anomalous sea level pressure field in the northern Pacific during spring after El Niño in models is stronger than that in observations, which is attributed to the excessive mean precipitation in the tropical Pacific in models. Kosaka and Nakamura³⁹ found that the ability of the models to reproduce the Pacific-Japan (PJ) pattern is significantly related to the simulated climatological mean state in the WNP. Models with smaller biases in WNP background fields are better able to reproduce the PJ pattern. Paek et al.⁴⁰ showed that the relative strength of the simulated mean Hadley and Walker circulations is critical to a realistic simulation of the summer WNPSH variability in the AGCM, in which the circulation causes different sensitivities to central Pacific (CP) and eastern Pacific (EP) El Niño in the model. The stronger mean Hadley circulation leads to the higher sensitivity of the model to CP El Niño, while the weaker mean Walker circulation leads to the lower sensitivity of the model to EP El Niño, which affects the simulation of the interannual variation of the Western Pacific subtropical high in the model.

Previous studies deepen our understanding of the causes and effects of the ENSO-related SSTA bias in coupled models, while there are still many uncertainties in understanding the impact of atmospheric mean state bias on the simulation bias of ENSO-related WNP anomalous circulation in coupled models. Considering coupled models play an important role in both climate research and prediction, it is necessary to examine the role of atmospheric mean state in modulating the simulation bias in coupled models. Motivated by these uncertainties, in the current study we intend to analyze the simulation biases of WNPAC among coupled models and the atmospheric mean state dependence during El Niño mature winter. The remaining part of this paper is organized as following. “Data and methods” section introduces the data and analysis methods. “Simulation biases of WNPAC in CMIP6 multi-model ensemble” section simply selects models and discusses simulation biases of WNPAC in CMIP6 multi-model ensemble. The bias of simulated WNPAC among coupled models and the relationship between the intermodel spread and different variables are discussed in “Origins of intermodel spread in CMIP6 models simulating WNPAC” section. Finally, a conclusion and discussions are given in “Summary and discussion” section.

Data and methods

Data

Monthly data of the historical simulations from 45 models of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experiment are utilized in the current study. The CMIP historical simulations begin with arbitrary equilibrium conditions from the pre-industrial control experiment (piControl) and integrate with time-dependent observational forcing due to anthropogenic changes in atmospheric composition (e.g., changes in greenhouse gases and aerosols) and land use, and as well as changes in natural forcing factors (e.g., changes in solar insolation and volcanic eruptions). The CMIP6 models included in this study are listed in Table 1, which range from 1850 to 2014. In addition, for comparison with the observational records in long-term perspective, the observed and reanalysis datasets for the same period (from 1850 to 2014) are used. Among them are monthly wind, temperature, specific humidity, precipitation and SST datasets from NOAA 20CR (NOAA/CIRES/DOE 20th Century Reanalysis V3) and monthly SST dataset from ERSST V5 (NOAA Extended Reconstructed SST V5) since 1854. The horizontal resolution of observed and reanalysis datasets above are both 1° × 1°, and all the observational and model data above are interpolated onto 2° × 2° grids. To remove the global warming trend, a least squares linear trend is applied to each field.

Table 1.

Model names, modeling centers and countries, and spatial resolutions of CMIP6 global climate models.

Model name	Modeling center and country	Spatial resolution (lon × lat)
ACCESS-CM2	CSIRO-ARCCSS, Australia	1.875° × 2.5°
ACCESS-ESM1-5	CSIRO, Australia	1.875° × 2.483°
AWI-CM-1-1-MR	AWI, Germany	0.938° × 1.875°
AWI-ESM-1-1-LR	AWI, Germany	1.875° × 3.75°
BCC-CSM2-MR	BCC, China	1.125° × 2.25°
BCC-ESM1	BCC, China	2.813° × 5.625°
CAMS-CSM1-0	CAMS, China	1.125° × 2.25°
CanESM5	CCCMA, Canada	2.813° × 5.625°
CESM2	NCAR, USA	1.25° × 1.875°
CESM2-FV2	NCAR, USA	2.5° × 3.75°
CESM2-WACCM	NCAR, USA	1.25° × 1.875°
CIESM	THU, China	1.25° × 1.875°
CMCC-CM2-HR4	CMCC, Italy	1.25° × 1.875°
CMCC-CM2-SR5	CMCC, Italy	1.25° × 1.875°
CMCC-ESM2	CMCC, Italy	1.25° × 1.875°
E3SM-1-0	E3SM-Project, USA	1° × 2°
E3SM-1-1	E3SM-Project, USA	1° × 2°
EC-Earth3	EC-Earth-Consortium, EU	0.703° × 1.406°
EC-Earth3-AerChem	EC-Earth-Consortium, EU	0.703° × 1.406°
EC-Earth3-CC	EC-Earth-Consortium, EU	0.703° × 1.406°
EC-Earth3-Veg	EC-Earth-Consortium, EU	0.703° × 1.406°
FGOALS-f3-L	CAS, China	1° × 2°
FGOALS-g3	CAS, China	2° × 2°
FIO-ESM-2-0	FIO-QLNM, China	1.25° × 1.875°
GFDL-CM4	NOAA GFDL, USA	1.25° × 2°
GFDL-ESM4	NOAA GFDL, USA	1.25° × 2°
GISS-E2-1-G	NASA-GISS, USA	2.5° × 4°
GISS-E2-1-H	NASA-GISS, USA	2.5° × 4°
GISS-E2-2-G	NASA-GISS, USA	2.5° × 4°
GISS-E2-2-H	NASA-GISS, USA	2.5° × 4°
IITM-ESM	CCCR-IITM, India	1.875° × 3.83°
IPSL-CM5A2-INCA	IPSL, France	3.75° × 3.75°
IPSL-CM6A-LR	IPSL, France	2.5° × 2.517°
KACE-1-0-G	NIMS-KMA, Korea	1.875° × 2.5°
KIOST-ESM	KIOST, Korea	1.875° × 3.83°
MIROC6	MIROC, Japan	1.406° × 2.813°
MPI-ESM-1-2-HAM	HAMMOZ-Consortium, EU	1.875° × 3.75°
MPI-ESM1-2-HR	MPI-M, Germany	0.938° × 1.875°
MPI-ESM1-2-LR	MPI-M, Germany	1.875° × 3.75°
MRI-ESM2-0	MRI, Japan	1.125° × 2.25°
NESM3	NUIST, China	1.875° × 3.75°
NorCPM1	NCC, Norway	2.5° × 3.75°
NorESM2-LM	NCC, Norway	2.5° × 3.75°
NorESM2-MM	NCC, Norway	1.25° × 1.875°
TaiESM1	AS-RCEC, Taiwan	1.25° × 1.875°

Open in a new tab

Methods

We identify El Niño and La Niña events during 1854–2014 according to the criterion that the Niño-3.4 SST index [SST anomalies averaged over 120°–170°W, 5°S–5°N]⁴¹ exceeds 1 standard deviation (σ) or falls below − 1 σ, respectively. Here we focus on evaluating the model performance during El Niño mature winter [December–February (DJF)].

Following Lorenz⁴², an empirical orthogonal function (EOF) analysis is used to obtain the leading modes of tropical Pacific SST during DJF. A linear regression method is used to investigate the intermodel relationship between the WNPAC intensity and DJF mean precipitation. To quantitatively measure the models’ performance in capturing the WNPAC strength, a WNPAC intensity index is introduced following He and Zhou⁴³ and Wang et al.⁴⁴. Each model is treated as a sample and the intermodel regression is done by regressing the DJF mean precipitation onto the WNPAC intensity index. The multi-model ensemble mean (MME) is calculated as an arithmetic average of variables among the 45 models.

A moisture budget analysis is used to verify the role of mean-state moisture biases and clarify the sources of anomalous precipitation differences between models presenting strong and weak WNPAC. The column-integrated moisture budget can be written as follows⁴⁵:

where angle brackets represent a mass-weighted vertical integral from 1000 to 100 hPa level, P is rainfall rate, E is the surface evaporation rate, q denotes the specific humidity, ∇ represents the horizontal gradient operator, Inline graphic is the two-dimensional wind vectors. Under climate state, may be approximated as 0 and omitted. When discussing the role of mean-state biases in affecting WNPAC biases, we separate each physical variable into climate state and anomaly⁴⁶, ie: , , .

Thus, the precipitation anomaly between models with WNPAC index greater than 1 standard deviation and models with WNPAC index less than − 1 standard deviation can be written as

where Inline graphic denotes the difference between models with WNPAC index greater than 1 standard deviation and models with WNPAC index less than − 1 standard deviation.

Simulation biases of WNPAC in CMIP6 multi-model ensemble

Previous studies found that the WNPAC during El Niño mature winter is the result of the anomalous heat source in the central-east Pacific through a two-step response^{4,10,14,19,20}. Therefore, it is necessary to examine whether models can reproduce El Niño events reasonably before discussing the impact of background fields on the WNPAC bias. In this section, we first evaluate the models’ performance and select models simply based on ERSST and NOAA 20CR SST data.

Figure S1a, b presents the SSTA patterns regressed onto the corresponding standardized Niño-3.4 SST index during boreal winter (DJF) for the period of 1854–2014 calculated by ERSST and for the period of 1850–2014 calculated by NOAA 20CR. It is clearly shown that there exists central-eastern Pacific warming with V-shaped cold signals on the northwest and southwest sides whether in ERSST or NOAA 20CR, which represents the typical characteristic of ENSO⁴⁷. The same method is utilized to examine the ability of models to reproduce the SSTA mode in DJF. Figure S2 show the SSTA patterns regressed onto the standardized Niño-3.4 SST index during DJF in all the CMIP6 historical climate experiments. The positive SSTA in the central-eastern Pacific is significantly underestimated in INM-CM4-8 and INM-CM5-0, and the pattern correlation coefficient is also significantly lower. In addition, the positive SSTA shifts towards the southern Hemisphere in MCM-UA-1-0, with a large position deviation. Therefore, it is considered that the three models cannot reasonably reproduce El Niño events, and will not be used in the following discussion.

Before discussing the intermodel spread, we compare the anomalous precipitation and low-level circulation during El Niño mature winter in CMIP6 models and NOAA 20CR Reanalysis datasets. It is clearly shown in Fig. 1a,b that the location of positive precipitation anomaly in the central Pacific and cyclonic anomalous circulation in the MME are both further west than those in NOAA 20CR, which is the result of excessive westward extension of the positive SSTA (Fig. 1c,d)^48,49. When the cyclonic anomalous circulation advances to the west of 160°E, the WNPAC has been squeezed further west. In addition, there is a strong anticyclonic circulation in the northern Indian Ocean in the MME (Fig. 1a), while the anticyclonic circulation in NOAA 20CR is weak in this region, which indicates that there exists an excessive westward extension of WNPAC in the models. The anomalous precipitation in the TIO in the MME shows an obvious east–west dipole distribution, while there is no such a feature in NOAA 20CR.

Fig. 1 — Composite patterns of precipitation anomalies (shading; mm day⁻¹) and 850-hPa wind anomalies (vectors; m s⁻¹) during El Niño mature winter derived from (a) MME mean, and (b) NOAA 20CR. (c) Same as (a), (d) same as (b), but for SST anomalies (shading; K). This figure is created by The NCAR Command Language (NCL) Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

Figure 2a–c present the horizontal distributions of 850-hPa anomalous relative vorticity in CMIP6 models and NOAA 20CR Reanalysis datasets. Here we introduce a WNPAC intensity index, which is defined as the opposite sign of area-averaged anomalous relative vorticity at 850 hPa during El Niño mature winter over a key analysis domain in the WNP. The domain selection is based on the horizontal distribution of the MME of the simulated 850-hPa relative vorticity field (Fig. 2a). Although there is strong anomalous negative vorticity in the northern Indian Ocean, which we focus on is the anomalous circulation in the WNP. Therefore, the key domain is located in the WNP to the South China Sea, which covers 2°–20°N, 105°–135°E. As is shown in Fig. 2d, except for three models with negative WNPAC index, the remaining models can all reproduce anomalous anticyclone, but the intermodel spread is large. The WNPAC intensity index from MME is close to that from NOAA 20CR, which indicates that the WNPAC intensity difference between observation and MME is little. It is clearly seen in Fig. 2c that the negative vorticity is weaker in the Western Pacific east of the Philippines, while stronger in the northern Indian Ocean in the MME. The differences in both regions present high intermodel consistency, further confirming that the location of WNPAC is to the west and extends excessively towards the Indian Ocean in the models as compared with observations.

Fig. 2 — Composite patterns of the relative vorticity anomalies (unit: s⁻¹) on 850 hPa for (a) the MME mean and (b) NOAA 20CR, and (c) their difference during El Niño mature winter. Regions in which 70% of the models have the same sign of relative vorticity anomalies are stippled. The dashed black box in (a) indicates the key region to define the WNPAC intensity index (2°–20°N, 105°–135°E). (d) The WNPAC intensity index derived from individual models in CMIP6. The open bar represents the MME, and the hatched bar represents the observational counterpart. This figure is created by NCL Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

The biases of WNPAC in the MME may be attributed to the combined effects of SSTA and atmospheric mean state bias. First, the westward shift of the positive SSTA in the models causes the westward shift of both the positive precipitation anomaly in the central Pacific and the stimulated cyclonic circulation, resulting in the westward shift of WNPAC. Second, as is shown in Fig. 3, the mean-state meridional negative vorticity gradient is smaller over the WNP in the MME, with a larger value of equivalent beta. According to the Rossby wave adjustment mechanism¹⁵, the Rossby wave response to the anomalous heat source under a larger equivalent beta obtains larger scale and intensity, which is favorable for the westward stretch of the cyclonic anomalies stimulated by the anomalous heat source in the central Pacific.

Fig. 3 — Composite patterns of the mean-state meridional vorticity gradient (unit: m⁻¹ s⁻¹) on 850 hPa for (a) the MME mean and (b) NOAA 20CR during winter. This figure is created by NCL Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

The excessive westward stretch of the WNPAC to the Indian Ocean in the MME may be attributed to the difference in the precipitation anomaly in the Indian Ocean. It can be seen in Fig. 1 that there is stronger negative precipitation anomaly in the eastern Indian Ocean in the models, which stimulates stronger anticyclonic Rossby wave response, and promotes the westward stretch of the anomalous anticyclone to the northern Indian Ocean. A further question is, what causes the difference in the precipitation anomaly? First, the positive SSTA over the western Indian Ocean is stronger in the MME (Fig. 1c), which stimulates stronger positive anomalous precipitation and upward motion in the western Indian Ocean, and causes stronger anomalous downward motion and negative precipitation anomaly in the eastern Indian Ocean through the zonal anomalous vertical circulation. Second, it is shown in Fig. 4 that there is higher mean-state precipitation over the equatorial Indian Ocean in the MME, which promotes the local circulation-convection feedback, and helps enhance the positive anomalous precipitation and the corresponding anomalous vertical circulation in the western Indian Ocean, resulting in stronger anomalous downward motion and negative precipitation anomaly in the eastern Indian Ocean.

Fig. 4 — Composite patterns of mean precipitation (shading; mm day⁻¹) and 850-hPa wind (vectors; m s⁻¹) for (a) the MME mean and (b) NOAA 20CR, and (c) their difference during winter. This figure is created by NCL Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

Origins of intermodel spread in CMIP6 models simulating WNPAC

Before discussing the role of each factor in the intermodel spread in simulating the WNPAC, we first examine the relationship between the intensity of WNPAC and the anomalous precipitation and low-level anomalous circulation. An intermodel regression among the 45 models is done by regressing the anomalous physical variable fields onto the standardized WNPAC indices. Figure 5 shows the results of intermodel regression and the comparison of model composite with WNPAC index greater than 1 standard deviation and less than − 1 standard deviation. It is clearly shown that the positive anomalous precipitation and cyclonic anomalous circulation in the central Pacific and the negative anomalous precipitation in the WNP are stronger in the models with stronger WNPAC. The associated process can be summarized as follows. The WNPAC can be viewed as a two-step response to the anomalous heating in the central Pacific caused by El Niño^4,14,19. In the first-step response, the anomalous heat source in the central Pacific caused by El Niño is stronger, stimulating stronger cyclonic anomalous wind field. The northeasterly anomalies to the northwest flank of the cyclone anomaly can enhance the negative precipitation anomaly in the WNP through wind-evaporation-SST feedback mechanism^4,19 and the atmospheric moist enthalpy advection process¹⁴, which promotes the second-step response and strengthen the WNPAC.

It is shown in Fig. 5b,c that there are significant differences in the intensity and location of anomalous circulation in models with high and low WNPAC indices. In models with a higher index, the cyclonic circulation in the central Pacific extends westward to 140°–150°E, and the WNPAC is located near the Philippines. However, in models with a lower index, the positive precipitation anomaly is weak and concentrated in the western Pacific, and the cyclonic circulation is also weak but extends farther westward to around 130°E. The anomalous anticyclone is almost invisible in the WNP, with only parts in the South China Sea and the northern Indian Ocean. In addition, the dipole pattern of precipitation anomaly over the Indian Ocean is stronger in the models with a higher index, which also presents in the models with a lower index, but with a weaker precipitation anomaly.

Different from AMIP experiments forced by prescribed historical SST and sea ice, historical experiments are air–sea coupling experiments. There are large intermodel differences in SSTA during El Niño⁵⁰. Therefore, the influence of SSTA on the intermodel spread of WNPAC cannot be ignored. Figure 6 shows the relationship between SSTA and WNPAC. It is clearly seen that the regression coefficient is positive in the equatorial central-eastern Pacific and off the coast of East Asia, and negative in the equatorial Western Pacific and WNP. Except for the equatorial Western Pacific, its modes in the Pacific region are basically consistent with the SSTA modes during El Niño mature winter (Fig. 1c). The results indicate that the models with stronger positive SSTA during El Niño and negative SSTA in the WNP tend to produce stronger WNPAC, which is also verified in the comparison of SSTA patterns between models with high and low WNPAC indices (Fig. 6b,c). In addition, the positive SSTA in the equatorial central Pacific in models with higher indices is stronger, but the westward extension is smaller than that in models with lower indices, which corresponds to the negative regression coefficient in the equatorial Western Pacific.

Fig. 6 — Same as Fig. 5, except for (a) regression and (b, c) composite of SST anomalies (unit: K). This figure is created by NCL Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

The relationship between SSTA and WNPAC intensity can be understood as follows. The stronger positive SSTA in the central-eastern Pacific causes stronger positive precipitation anomaly and heating anomaly, and stimulates stronger cyclonic anomalous circulation. The cyclone anomaly causes stronger negative precipitation anomaly and eventually stimulates stronger anomalous anticyclone in the WNP through wind-evaporation-SST feedback^4,19, and moist enthalpy advection mechanism¹⁴. In regard to the difference in location, a smaller westward extension of positive SSTA corresponds to the more eastward positive precipitation anomaly and cyclonic circulation, therefore, the WNPAC could develop to the east of the Philippines. Furthermore, the anomalous anticyclones in the northern Indian Ocean are also stronger in models with higher indices. A possible cause is that the stronger positive SSTA in the western Indian Ocean tends to cause a stronger local positive precipitation anomaly, and strengthen the negative precipitation anomaly in the eastern Indian Ocean through the zonal anomalous vertical circulation, which is in favor of the extension of WNPAC to the northern Indian Ocean.

The above discussion indicates that the uncertainty of WNPAC can be explained by the uncertainty of SSTA associated with El Niño in coupled models through specific physical processes. Is it still necessary to focus on the atmospheric background fields? The previous analysis of AMIP experiments revealed the importance of mean-state precipitation/moisture, vorticity and stability on the uncertainty of WNPAC⁴⁴. Here we also focus on these three factors and discuss the relative roles in coupling experiments.

Figure 7 shows the relationship between mean-state precipitation, 850-hPa wind field, low-level specific humidity and the WNPAC index during El Niño mature winter. In models with stronger WNPAC, there is more mean precipitation in the equatorial Pacific and the coastal region of the western Indian Ocean, but less mean precipitation in the southern equatorial region of the Indian Ocean. The distribution of mean-state moisture in the equatorial Pacific and Indian Ocean is basically consistent with that of mean-state precipitation, except for the location of largest value in the equatorial Pacific and the insignificant negative signal to the south of equatorial Indian Ocean. The vertical profile of specific humidity (Fig. 7c) shows that there is a significant signal from low level to middle level in the equatorial Pacific.

The relationship between mean precipitation in the equatorial Pacific (red box in Fig. 7) and the WNPAC intensity can be further confirmed by the scatter plot. Figure 8 also shows the relationship between SSTA over the WNP (0°–20°N, 130°–160°E; green box in Fig. 6a), SSTA over the central eastern Pacific (Niño-3.4 region), mean SST over the equatorial Pacific and the WNPAC intensity. The vertical axis is the WNPAC index, and the horizontal axes are the above factors. The correlation between SSTA over the WNP and the WNPAC intensity is strongest, which is because SSTA in this region can directly affect the anomalous circulation in the WNP by stimulating local anomalous precipitation. SSTA in Niño-3.4 region and mean precipitation in the equatorial Pacific are significantly correlated with the WNPAC index. The correlation coefficients are 0.61 and 0.60. We also calculated the partial correlation coefficient among these three variables. The partial correlation coefficient between SSTA and WNPAC index is 0.57 when the influence of mean precipitation is excluded, while the partial correlation coefficient between mean precipitation and WNPAC index is 0.56 when the influence of SSTA is excluded. There are little differences with the directly calculated correlation coefficients, which indicates that the contribution of mean precipitation in the equatorial Pacific and SSTA in the Niño-3.4 region to the difference of WNPAC is relatively independent. Even though the difference of SSTA is considered, the difference of mean precipitation is still important in coupling experiments. Although the mean SST in coupling experiments can explain the difference of mean precipitation in some extent, the correlation between mean SST in the equatorial Pacific and the WNPAC index is low, which further indicates that it is necessary to focus on the atmospheric mean state when discussing the source of WNPAC uncertainty in coupled models, and it is not comprehensive to focus only on the difference of SST.

Fig. 8 — Scatter diagrams between the multiple factors and the WNPAC index: averaged SST anomalies over (a) the western North Pacific [inside the left green-outlined box in Fig. 6a] and (b) the central eastern Pacific [inside the right green-outlined box in Fig. 6a, i.e. Niño-3.4 region]; (c) mean precipitation and (d) mean SST over the equatorial Pacific [inside the red-outlined box in Fig. 7]. The correlation coefficients in (a–c) exceed the 99% confidence level, and the correlation coefficient in (d) exceeds the 95% confidence level.

How does the mean-state precipitation/moisture field affect the uncertainty of WNPAC? The modulation of mean-state precipitation in AMIP experiments can be explained by a circulation-convection feedback⁴⁴. That is, the initial perturbation with same intensity would induce stronger moisture convergence and enhance precipitation in the mean-state with larger precipitation/moisture. Then, the initial perturbation would develop into stronger anomalous circulation and further enhance moisture convergence and precipitation. Therefore, atmospheric circulation in the models with larger mean-state precipitation/moisture have higher sensitivity to SSTA. Can a similar mechanism be used for coupling experiments? For addressing the above issues, we use the regression coefficient between anomalous precipitation and local SSTA to describe the sensitivity of atmosphere to SSTA in models. The intensity of circulation-convection feedback is described through the following method: The calculation region is located in the equatorial Pacific (red box in Fig. 7). By regressing the DJF anomalous precipitation from 1850 to 2014 onto SSTA in the same period, the larger regression coefficient represents that the same SSTA forcing can stimulate stronger anomalous precipitation, indicating that the corresponding model is more sensitive to SSTA on the interannual time scale.

Figure 9 shows the regression coefficients representing the sensitivity of each model to SSTA (vertical axis in Fig. 9a, horizontal axis in Fig. 9b) and the relationship between the regression coefficients and the mean precipitation/the WNPAC index (horizontal axis in Fig. 9a/vertical axis in Fig. 9b). Note that there is a strong positive correlation between the sensitivity of models to SSTA and the mean precipitation. And the models with higher sensitivity tend to induce a stronger WNPAC. The physical process can also be explained through the circulation-convection feedback mechanism. The positive SSTA stimulates positive anomalous precipitation and the corresponding anomalous circulations such as low-level convergence, ascending motion and upper-level divergence. Assume there is more mean precipitation and moisture. The anomalous low-level convergence and ascending motion transport more moisture upward, causing more anomalous precipitation and releasing more latent heat of condensation, which further strengthens anomalous circulation. In other words, the greater mean-state precipitation and moisture strengthen the local circulation-convection feedback, which promotes the sensitivity of models to SSTA and stimulates stronger anomalous precipitation with the same SSTA intensity, then the first-step response caused by El Niño is strengthened. The strengthened cyclonic circulation enhances the second-step response in the WNP, and induces a stronger WNPAC.

In order to further verify the role of mean-state moisture biases and clarify the sources of anomalous precipitation differences between models presenting strong and weak WNPAC, we also perform a vertical (1000–100 hPa) integrated moisture budget diagnosis. Figure 10 displays the contribution from each of the moisture budget terms. It is found that the term of vertically integrated horizontal convergence of the mean moisture by the anomalous wind (i.e., Inline graphic ) plays a dominant role in the anomalous precipitation differences. Also, we further decompose this term and separate the mean moisture and anomalous wind into MME-averaged component and the deviation from it. It is indicated that the leading term comes from vertically integrated horizontal convergence of the MME-averaged moisture by the anomalous wind of deviation (Figure not shown). However, the anomalous wind bias may be influenced by the mean-state moisture or precipitation bias. It is hard to separate the effects of mean moisture and anomalous wind, which also reflects the limitation of moisture budget diagnosis in our research.

Fig. 10 — Vertically integrated (1000–100 hPa) moisture budget terms (unit: mm day⁻¹) of difference in precipitation anomaly between models with WNPAC index greater than 1 standard deviation and models with WNPAC index less than − 1 standard deviation averaged over region (6°S–6°N, 160°E–120°W) [inside the red-outlined box in Fig. S4].

Besides mean-state precipitation and moisture, there is also a significant relationship between the mean-state vorticity and the uncertainty of WNPAC. In models with high WNPAC indices, the positive vorticity to the north of central western equatorial Pacific is stronger, therefore, the negative vorticity gradient is larger in the WNP (Fig. 11). A larger negative meridional vorticity gradient represents a smaller equivalent beta, and the cyclonic circulation stimulated by central Pacific heating extends less westward through the Rossby wave adjustment mechanism, which leaves more space for the development of the anomalous anticyclone. An opposite process operates in models with low WNPAC indices. The larger equivalent beta causes that the location of cyclonic circulation is to the west, resulting in the anomalous anticyclone confined in the South China Sea and the northern Indian Ocean. The impact of mean-state vorticity on anomalous circulation has been verified by ideal numerical experiments in the previous studies¹⁵.

Fig. 11 — (a) The DJF mean 850-hPa relative vorticity fields (s⁻¹) regressed onto the standardized WNPAC indices. Stippled regions indicate that the regressed mean vorticity exceeds the 95% confidence level. (b) Composite patterns of DJF mean 850-hPa meridional vorticity gradient (m⁻¹ s⁻¹) for models with WNPAC index greater than 1 standard deviation, (c) same as (b), but for models with WNPAC index less than − 1 standard deviation. This figure is created by NCL Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

The mean-state stability contributes little to the uncertainty of WNPAC in the coupling experiments. Although the regression results show that the models with stronger WNPAC have lower mean-state stability over a large area from the tropical Pacific to the Indian Ocean, the relationship is not significant in the WNP, the South China Sea and the northern Indian Ocean where the anomalous anticyclone circulation is located (Fig. 12).

Fig. 12 — The DJF mean middle-level (600–400-hPa average) static stability fields (K Pa⁻¹) regressed onto the standardized WNPAC indices. Stippled regions indicate that the regressed static stability exceeds the 95% confidence level. This figure is created by NCL Version 6.5.0 (URL: http://dx.doi.org/10.5065/D6WD3XH5).

Summary and discussion

In this study, we utilized the CMIP6 historical experiment datasets to study the influence of atmospheric background fields on the intermodel spread of WNPAC during El Niño mature winter, and verify that both the modulation of background fields and SSTA biases play important roles in the coupling experiments.

Most of the models are capable of reproducing the observed pattern of SSTA during El Niño mature winter, but the excessive westward extension of positive SSTA during El Niño is widespread, which results in the westward shift of both the positive precipitation anomaly and the cyclonic circulation in the central Pacific in the MME, and then the westward shift of WNPAC. Moreover, the mean-state meridional negative vorticity gradient in the WNP in the MME is smaller, which is also conducive to the westward development of cyclonic anomalous circulation stimulated by the anomalous heat source in the Central Pacific through the Rossby wave adjustment mechanism. In addition, there is a problem of excessive extension of WNPAC to the Indian Ocean in the MME, which is attributed to the combined effects of SSTA and mean precipitation biases. The stronger positive SSTA in the western Indian Ocean and the larger mean precipitation in the equatorial Indian Ocean in the MME can both enhance the positive precipitation anomaly in the western Indian Ocean, and cause the enhancement of negative precipitation anomaly in the eastern Indian Ocean through the zonal anomalous vertical circulation, which is favorable for the westward stretch of the anomalous anticyclone to the northern Indian Ocean.

The intensity and zonal location of WNPAC in historical experiments present large intermodel spread. The sources of intermodel spread were explored through intermodel regression in this study. Firstly, the intermodel difference in WNPAC intensity is attributed to the difference in anomalous precipitation in the WNP, which comes from the difference in anomalous heat source and cyclonic anomalous circulation in the central Pacific, that is, the difference in the first-step response caused by El Niño. The uncertainty of anomalous precipitation and circulation can be further attributed to the SSTA in the Niño-3.4 region and the mean-state precipitation/moisture in the equatorial Pacific. On the one hand, the intensity of positive SSTA can directly affect the intensity of positive anomalous precipitation induced by it. On the other hand, the amount of mean-state precipitation/moisture determines the intensity of local circulation-convective feedback, resulting in different sensitivities of atmosphere to SSTA in models, which can also modulate the intensity of anomalous precipitation induced by SSTA. The modulation of the two is relatively independent. In regard to the difference in the zonal location of WNPAC, except for the influence of the westward extension of positive SSTA during El Niño, the mean-state vorticity also makes contribution, that is, the difference of mean meridional vorticity gradient in the WNP modulates the westward extension of cyclonic circulation through the Rossby wave adjustment mechanism, and then affects the zonal location of WNPAC.

The conclusion above illustrates that the atmospheric background fields play an important role in modulating the intermodel spread of WNPAC in coupled models. It is worthy to compare the relative contributions of SSTA difference and atmospheric background mean state difference through numerical experiments in future endeavors. In addition, the reason for intermodel spread of atmospheric background fields remains unknown, it is necessary to explore the sources of background field differences based on the physical processes in models in the future.

Supplementary Information

Supplementary Information.^{(28.7MB, docx)}

Acknowledgements

We are grateful to the public data websites for providing the datasets used in this study at no cost.

Author contributions

JL and XW designed and performed the research and analyzed the data. JL drafted the manuscript, XW, YG, ZC, ZH and XC provided the comments and revised the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (42305042), Jiangsu Meteorological Bureau Scientific Research Project (KQ202512), Suzhou Meteorological Bureau Scientific Research Project (SZKJ202303), Suzhou Science & Technology Planning Project (2023ss10), Jiangsu Natural Science Key Project (23KJB170004), and the Open Foundation of China Meteorological Administration Hydro-Meteorology Key Laboratory (23SWQXM008).

Data availability

All the dataset adopted in this study can be accessed online via the following URL. The CMIP6 model data: https://esgf-node.llnl.gov/projects/cmip6. NOAA/CIRES/DOE 20th Century Reanalysis version 3: https://psl.noaa.gov/data/gridded/data.20thC_ReanV3.html. Extended Reconstructed SST version 5: https://psl.noaa.gov/data/gridded/data.noaa.ersst.v5.html.

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Jiahao Lu, Email: nuistljh@126.com.

Xieyuan Wang, Email: 18252086395@163.com.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-04048-7.

References

1.Zhang, R. H., Sumi, A. & Kimoto, M. Impact of El Niño on the East Asian monsoon: A diagnostic study of the ’86/87 and ’91/92 events. J. Meteor. Soc. Japan.74, 49–62 (1996). [Google Scholar]
2.Chang, C.-P., Zhang, Y. S. & Li, T. Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part I: Roles of the subtropical ridge. J. Clim.13, 4310–4325 (2000). [Google Scholar]
3.Chang, C.-P., Zhang, Y. S. & Li, T. Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part II: Meridional structure of the monsoon. J. Clim.13, 4326–4340 (2000). [Google Scholar]
4.Wang, B., Wu, R. G. & Fu, X. H. Pacific-East Asian teleconnection: How does ENSO affect East Asian climate?. J. Clim.13, 1517–1536 (2000). [Google Scholar]
5.Wu, R. G., Hu, Z. Z. & Kirtman, B. P. Evolution of ENSO-related rainfall anomalies in East Asia. J. Clim.16, 3742–3758 (2003). [Google Scholar]
6.Xie, S.-P. et al. Indo-western Pacific Ocean capacitor and coherent climate anomalies in post-ENSO summer: A review. Adv. Atmos. Sci.33, 411–432 (2016). [Google Scholar]
7.Li, T. & Hsu, P.-C. Monsoon dynamics and its interactions with ocean. In: Fundamentals of Tropical Climate Dynamics, 236 (Springer, 2017).
8.Okumura, Y. M., Ohba, M., Deser, C. & Ueda, H. A proposed mechanism for the asymmetry duration of El Niño and La Niña. J. Clim.24, 3822–3829 (2011). [Google Scholar]
9.Chen, M.-C., Li, T., Shen, X.-Y. & Wu, B. Relative roles of dynamic and thermodynamic processes in causing evolution asymmetry between El Niño and La Niña. J. Clim.29, 2201–2220 (2016). [Google Scholar]
10.Wang, B., Wu, R. G. & Li, T. Atmosphere–warm ocean interaction and its impacts on Asian-Australian monsoon variation. J. Clim.16, 1195–1211 (2003). [Google Scholar]
11.Xie, S.-P. et al. Indian Ocean capacitor effect on Indo-western Pacific climate during the summer following El Niño. J. Clim.22, 730–747 (2009). [Google Scholar]
12.Wu, B., Zhou, T. & Li, T. Seasonally evolving dominant interannual variability modes of East Asian climate. J. Clim.22, 2992–3005 (2009). [Google Scholar]
13.Wu, B., Li, T. & Zhou, T. Relative contributions of the Indian Ocean and local SST anomalies to the maintenance of the western North Pacific anomalous anticyclone during the El Niño decaying summer. J. Clim.23, 2974–2986 (2010). [Google Scholar]
14.Wu, B., Zhou, T. & Li, T. Atmospheric dynamic and thermodynamic processes driving the western North Pacific anomalous anticyclone during El Niño. Part I Maint. Mech. J. Clim.30, 9621–9635 (2017). [Google Scholar]
15.Wu, B., Zhou, T. & Li, T. Atmospheric dynamic and thermodynamic processes driving the western North Pacific anomalous anticyclone during El Niño. Part II Form. Process. J. Clim.30, 9637–9650 (2017). [Google Scholar]
16.Wang, B., Xiang, B. Q. & Lee, J.-Y. Subtropical high predictability establishes a promising way for monsoon and tropical storm predictions. Proc. Natl. Acad. Sci. USA110, 2718–2722 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Stuecker, M. F. et al. Combination mode dynamics of the anomalous Northwest Pacific anticyclone. J. Clim.28, 1093–1111 (2015). [Google Scholar]
18.Xiang, B. Q. et al. How can anomalous western North Pacific subtropical high intensify in late summer?. Geophys. Res. Lett.40, 2349–2354 (2013). [Google Scholar]
19.Wang, B. & Zhang, Q. Pacific-East Asian teleconnection. Part II: How the Philippine Sea anomalous anticyclone is established during El Niño development. J. Clim.15, 3252–3265 (2002). [Google Scholar]
20.Lau, N. & Nath, M. J. Atmosphere–ocean variations in the Indo-Pacific sector during ENSO episodes. J. Clim.16, 3–20 (2003). [Google Scholar]
21.Wang, C. & Picaut, J. Understanding ENSO physics-a review. In: Earth’s Climate: The Ocean-Atmosphere, 21–48 (American Geophysical Union, 2004).
22.Guilyardi, E. et al. Understanding El Niño in ocean-atmosphere general circulation models: Progress and challenges. Bull. Am. Meteorol. Soc.90, 325–340 (2009). [Google Scholar]
23.Bellenger, H. et al. ENSO representation in climate models: from CMIP3 to CMIP5. Clim. Dyn.42, 1999–2018 (2014). [Google Scholar]
24.Collins, M. et al. The impact of global warming on the tropical Pacific Ocean and El Niño. Nat. Geosci.3, 391–397 (2010). [Google Scholar]
25.Kim, S. T. & Yu, J.-Y. The two types of ENSO in CMIP5 models. Geophys. Res. Lett.39, L11704 (2012). [Google Scholar]
26.Kug, J.-S., Ham, Y.-G., Lee, J.-Y. & Jin, F.-F. Improved simulation of two types of El Niño in CMIP5 models. Environ. Res. Lett.7, 034002 (2012). [Google Scholar]
27.Jiang, W. et al. Biases and improvements of the ENSO-East Asian winter monsoon teleconnection in CMIP5 and CMIP6 models. Clim. Dyn.59, 2467–2480 (2022). [Google Scholar]
28.Gong, H. et al. Diverse influences of ENSO on the East Asian-western Pacific winter climate tied to different ENSO properties in CMIP5 models. J. Clim.28, 2187–2202 (2015). [Google Scholar]
29.Wang, Z. et al. What determine the performance of the ENSO-East Asian winter monsoon relationship in CMIP6 models?. J. Geophys. Res. Atmos.127, e2021JD036227 (2022). [Google Scholar]
30.Jiang, W. et al. Diverse relationship between ENSO and the northwest Pacific summer climate among CMIP5 models: Dependence on the ENSO decay pace. J. Clim.30, 109–127 (2017). [Google Scholar]
31.Li, G. et al. Effect of excessive equatorial Pacific cold tongue bias on the El Niño-northwest Pacific summer monsoon relationship in CMIP5 multi-model ensemble. Clim. Dyn.52, 6195–6212 (2019). [Google Scholar]
32.Jiang, W., Huang, P., Huang, G. & Ying, J. Origins of the excessive westward extension of ENSO SST simulated in CMIP5 and CMIP6 models. J. Clim.34, 2839–2851 (2021). [Google Scholar]
33.Bayr, T. et al. Mean-state dependence of ENSO atmospheric feedbacks in climate models. Clim. Dyn.50, 3171–3194 (2018). [Google Scholar]
34.Bayr, T. et al. Error compensation of ENSO atmospheric feedbacks in climate models and its influence on simulated ENSO dynamics. Clim. Dyn.53, 155–172 (2019). [Google Scholar]
35.Li, R. K. K., Tam, C. Y. & Lau, N. C. Effects of ENSO diversity and cold tongue bias on seasonal prediction of South China late spring rainfall. Clim. Dyn.57, 577–591 (2021). [Google Scholar]
36.Tao, W. et al. Dominant modes of CMIP3/5 models simulating northwest Pacific circulation anomalies during post-ENSO summer and their SST dependence. Theor. Appl. Climatol.138, 1809–1820 (2019). [Google Scholar]
37.Feng, J. et al. An investigation of CMIP5 model biases in simulating the impacts of central Pacific El Niño on the East Asian summer monsoon. Clim. Dyn.52, 2631–2646 (2019). [Google Scholar]
38.Spencer, H. & Slingo, J. M. The simulation of peak and delayed ENSO teleconnections. J. Clim.16, 1757–1774 (2003). [Google Scholar]
39.Kosaka, Y. & Nakamura, H. Dominant mode of climate variability, intermodel diversity, and projected future changes over the summertime Western North Pacific simulated in the CMIP3 models. J. Clim.24, 3935–3955 (2011). [Google Scholar]
40.Paek, H. et al. A source of AGCM bias in simulating the Western Pacific subtropical high: Different sensitivities to the two types of ENSO. Mon. Weather Rev.143, 2348–2362 (2015). [Google Scholar]
41.Barnston, A. G., Chelliah, M. & Goldenberg, S. B. Documentation of a highly ENSO-related SST region in the equatorial Pacific. Atmos. Ocean.35, 367–383 (1997). [Google Scholar]
42.Lorenz, E. N. Empirical orthogonal functions and statistical weather prediction 49 (Scientific Rep 1, Statistical forecasting project, Massachusetts Institute of Technology, Department of Meteorology, 1956).
43.He, C. & Zhou, T. The two interannual variability modes of the western North Pacific subtropical high simulated by 28 CMIP5-AMIP models. Clim. Dyn.43, 2455–2469 (2014). [Google Scholar]
44.Wang, X., Li, T. & He, C. Impact of the mean state on El Niño-induced western North Pacific anomalous anticyclones during El Niño decaying summer in AMIP models. J. Clim.34, 9201–9217 (2021). [Google Scholar]
45.Banacos, P. C. & Schultz, D. M. The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Weather Forecast.20, 351–366 (2005). [Google Scholar]
46.Peng, D. D. & Zhou, T. J. Why was the arid and semiarid northwest China getting wetter in the recent decades?. J. Geophys. Res. Atmos.122, 9060–9075 (2017). [Google Scholar]
47.Takahashi, K., Montecinos, A., Goubanova, K. & Dewitte, B. ENSO regimes: Reinterpreting the canonical and Modoki El Niño. Geophys. Res. Lett.38, L10704 (2011). [Google Scholar]
48.Misra, V. et al. Validating and understanding the ENSO simulation in two coupled climate models. Tellus59, 292–308 (2007). [Google Scholar]
49.Ham, Y.-G. & Kug, J.-S. Improvement of ENSO simulation based on intermodel diversity. J. Clim.28, 998–1015 (2015). [Google Scholar]
50.Deser, C., Simpson, I. R., McKinnon, K. A. & Phillips, A. S. The Northern hemisphere extratropical atmospheric circulation response to ENSO: How well do we know it and how do we evaluate models accordingly?. J. Clim.30, 5059–5082 (2017). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information.^{(28.7MB, docx)}

Data Availability Statement

[CR1] 1.Zhang, R. H., Sumi, A. & Kimoto, M. Impact of El Niño on the East Asian monsoon: A diagnostic study of the ’86/87 and ’91/92 events. J. Meteor. Soc. Japan.74, 49–62 (1996). [Google Scholar]

[CR2] 2.Chang, C.-P., Zhang, Y. S. & Li, T. Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part I: Roles of the subtropical ridge. J. Clim.13, 4310–4325 (2000). [Google Scholar]

[CR3] 3.Chang, C.-P., Zhang, Y. S. & Li, T. Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part II: Meridional structure of the monsoon. J. Clim.13, 4326–4340 (2000). [Google Scholar]

[CR4] 4.Wang, B., Wu, R. G. & Fu, X. H. Pacific-East Asian teleconnection: How does ENSO affect East Asian climate?. J. Clim.13, 1517–1536 (2000). [Google Scholar]

[CR5] 5.Wu, R. G., Hu, Z. Z. & Kirtman, B. P. Evolution of ENSO-related rainfall anomalies in East Asia. J. Clim.16, 3742–3758 (2003). [Google Scholar]

[CR6] 6.Xie, S.-P. et al. Indo-western Pacific Ocean capacitor and coherent climate anomalies in post-ENSO summer: A review. Adv. Atmos. Sci.33, 411–432 (2016). [Google Scholar]

[CR7] 7.Li, T. & Hsu, P.-C. Monsoon dynamics and its interactions with ocean. In: Fundamentals of Tropical Climate Dynamics, 236 (Springer, 2017).

[CR8] 8.Okumura, Y. M., Ohba, M., Deser, C. & Ueda, H. A proposed mechanism for the asymmetry duration of El Niño and La Niña. J. Clim.24, 3822–3829 (2011). [Google Scholar]

[CR9] 9.Chen, M.-C., Li, T., Shen, X.-Y. & Wu, B. Relative roles of dynamic and thermodynamic processes in causing evolution asymmetry between El Niño and La Niña. J. Clim.29, 2201–2220 (2016). [Google Scholar]

[CR10] 10.Wang, B., Wu, R. G. & Li, T. Atmosphere–warm ocean interaction and its impacts on Asian-Australian monsoon variation. J. Clim.16, 1195–1211 (2003). [Google Scholar]

[CR11] 11.Xie, S.-P. et al. Indian Ocean capacitor effect on Indo-western Pacific climate during the summer following El Niño. J. Clim.22, 730–747 (2009). [Google Scholar]

[CR12] 12.Wu, B., Zhou, T. & Li, T. Seasonally evolving dominant interannual variability modes of East Asian climate. J. Clim.22, 2992–3005 (2009). [Google Scholar]

[CR13] 13.Wu, B., Li, T. & Zhou, T. Relative contributions of the Indian Ocean and local SST anomalies to the maintenance of the western North Pacific anomalous anticyclone during the El Niño decaying summer. J. Clim.23, 2974–2986 (2010). [Google Scholar]

[CR14] 14.Wu, B., Zhou, T. & Li, T. Atmospheric dynamic and thermodynamic processes driving the western North Pacific anomalous anticyclone during El Niño. Part I Maint. Mech. J. Clim.30, 9621–9635 (2017). [Google Scholar]

[CR15] 15.Wu, B., Zhou, T. & Li, T. Atmospheric dynamic and thermodynamic processes driving the western North Pacific anomalous anticyclone during El Niño. Part II Form. Process. J. Clim.30, 9637–9650 (2017). [Google Scholar]

[CR16] 16.Wang, B., Xiang, B. Q. & Lee, J.-Y. Subtropical high predictability establishes a promising way for monsoon and tropical storm predictions. Proc. Natl. Acad. Sci. USA110, 2718–2722 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Stuecker, M. F. et al. Combination mode dynamics of the anomalous Northwest Pacific anticyclone. J. Clim.28, 1093–1111 (2015). [Google Scholar]

[CR18] 18.Xiang, B. Q. et al. How can anomalous western North Pacific subtropical high intensify in late summer?. Geophys. Res. Lett.40, 2349–2354 (2013). [Google Scholar]

[CR19] 19.Wang, B. & Zhang, Q. Pacific-East Asian teleconnection. Part II: How the Philippine Sea anomalous anticyclone is established during El Niño development. J. Clim.15, 3252–3265 (2002). [Google Scholar]

[CR20] 20.Lau, N. & Nath, M. J. Atmosphere–ocean variations in the Indo-Pacific sector during ENSO episodes. J. Clim.16, 3–20 (2003). [Google Scholar]

[CR21] 21.Wang, C. & Picaut, J. Understanding ENSO physics-a review. In: Earth’s Climate: The Ocean-Atmosphere, 21–48 (American Geophysical Union, 2004).

[CR22] 22.Guilyardi, E. et al. Understanding El Niño in ocean-atmosphere general circulation models: Progress and challenges. Bull. Am. Meteorol. Soc.90, 325–340 (2009). [Google Scholar]

[CR23] 23.Bellenger, H. et al. ENSO representation in climate models: from CMIP3 to CMIP5. Clim. Dyn.42, 1999–2018 (2014). [Google Scholar]

[CR24] 24.Collins, M. et al. The impact of global warming on the tropical Pacific Ocean and El Niño. Nat. Geosci.3, 391–397 (2010). [Google Scholar]

[CR25] 25.Kim, S. T. & Yu, J.-Y. The two types of ENSO in CMIP5 models. Geophys. Res. Lett.39, L11704 (2012). [Google Scholar]

[CR26] 26.Kug, J.-S., Ham, Y.-G., Lee, J.-Y. & Jin, F.-F. Improved simulation of two types of El Niño in CMIP5 models. Environ. Res. Lett.7, 034002 (2012). [Google Scholar]

[CR27] 27.Jiang, W. et al. Biases and improvements of the ENSO-East Asian winter monsoon teleconnection in CMIP5 and CMIP6 models. Clim. Dyn.59, 2467–2480 (2022). [Google Scholar]

[CR28] 28.Gong, H. et al. Diverse influences of ENSO on the East Asian-western Pacific winter climate tied to different ENSO properties in CMIP5 models. J. Clim.28, 2187–2202 (2015). [Google Scholar]

[CR29] 29.Wang, Z. et al. What determine the performance of the ENSO-East Asian winter monsoon relationship in CMIP6 models?. J. Geophys. Res. Atmos.127, e2021JD036227 (2022). [Google Scholar]

[CR30] 30.Jiang, W. et al. Diverse relationship between ENSO and the northwest Pacific summer climate among CMIP5 models: Dependence on the ENSO decay pace. J. Clim.30, 109–127 (2017). [Google Scholar]

[CR31] 31.Li, G. et al. Effect of excessive equatorial Pacific cold tongue bias on the El Niño-northwest Pacific summer monsoon relationship in CMIP5 multi-model ensemble. Clim. Dyn.52, 6195–6212 (2019). [Google Scholar]

[CR32] 32.Jiang, W., Huang, P., Huang, G. & Ying, J. Origins of the excessive westward extension of ENSO SST simulated in CMIP5 and CMIP6 models. J. Clim.34, 2839–2851 (2021). [Google Scholar]

[CR33] 33.Bayr, T. et al. Mean-state dependence of ENSO atmospheric feedbacks in climate models. Clim. Dyn.50, 3171–3194 (2018). [Google Scholar]

[CR34] 34.Bayr, T. et al. Error compensation of ENSO atmospheric feedbacks in climate models and its influence on simulated ENSO dynamics. Clim. Dyn.53, 155–172 (2019). [Google Scholar]

[CR35] 35.Li, R. K. K., Tam, C. Y. & Lau, N. C. Effects of ENSO diversity and cold tongue bias on seasonal prediction of South China late spring rainfall. Clim. Dyn.57, 577–591 (2021). [Google Scholar]

[CR36] 36.Tao, W. et al. Dominant modes of CMIP3/5 models simulating northwest Pacific circulation anomalies during post-ENSO summer and their SST dependence. Theor. Appl. Climatol.138, 1809–1820 (2019). [Google Scholar]

[CR37] 37.Feng, J. et al. An investigation of CMIP5 model biases in simulating the impacts of central Pacific El Niño on the East Asian summer monsoon. Clim. Dyn.52, 2631–2646 (2019). [Google Scholar]

[CR38] 38.Spencer, H. & Slingo, J. M. The simulation of peak and delayed ENSO teleconnections. J. Clim.16, 1757–1774 (2003). [Google Scholar]

[CR39] 39.Kosaka, Y. & Nakamura, H. Dominant mode of climate variability, intermodel diversity, and projected future changes over the summertime Western North Pacific simulated in the CMIP3 models. J. Clim.24, 3935–3955 (2011). [Google Scholar]

[CR40] 40.Paek, H. et al. A source of AGCM bias in simulating the Western Pacific subtropical high: Different sensitivities to the two types of ENSO. Mon. Weather Rev.143, 2348–2362 (2015). [Google Scholar]

[CR41] 41.Barnston, A. G., Chelliah, M. & Goldenberg, S. B. Documentation of a highly ENSO-related SST region in the equatorial Pacific. Atmos. Ocean.35, 367–383 (1997). [Google Scholar]

[CR42] 42.Lorenz, E. N. Empirical orthogonal functions and statistical weather prediction 49 (Scientific Rep 1, Statistical forecasting project, Massachusetts Institute of Technology, Department of Meteorology, 1956).

[CR43] 43.He, C. & Zhou, T. The two interannual variability modes of the western North Pacific subtropical high simulated by 28 CMIP5-AMIP models. Clim. Dyn.43, 2455–2469 (2014). [Google Scholar]

[CR44] 44.Wang, X., Li, T. & He, C. Impact of the mean state on El Niño-induced western North Pacific anomalous anticyclones during El Niño decaying summer in AMIP models. J. Clim.34, 9201–9217 (2021). [Google Scholar]

[CR45] 45.Banacos, P. C. & Schultz, D. M. The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Weather Forecast.20, 351–366 (2005). [Google Scholar]

[CR46] 46.Peng, D. D. & Zhou, T. J. Why was the arid and semiarid northwest China getting wetter in the recent decades?. J. Geophys. Res. Atmos.122, 9060–9075 (2017). [Google Scholar]

[CR47] 47.Takahashi, K., Montecinos, A., Goubanova, K. & Dewitte, B. ENSO regimes: Reinterpreting the canonical and Modoki El Niño. Geophys. Res. Lett.38, L10704 (2011). [Google Scholar]

[CR48] 48.Misra, V. et al. Validating and understanding the ENSO simulation in two coupled climate models. Tellus59, 292–308 (2007). [Google Scholar]

[CR49] 49.Ham, Y.-G. & Kug, J.-S. Improvement of ENSO simulation based on intermodel diversity. J. Clim.28, 998–1015 (2015). [Google Scholar]

[CR50] 50.Deser, C., Simpson, I. R., McKinnon, K. A. & Phillips, A. S. The Northern hemisphere extratropical atmospheric circulation response to ENSO: How well do we know it and how do we evaluate models accordingly?. J. Clim.30, 5059–5082 (2017). [Google Scholar]

PERMALINK

Impact of mean state on intermodel spread of western North Pacific anomalous anticyclone during El Niño mature winter in CMIP6 models

Jiahao Lu

Xieyuan Wang

Yuhan Gong

Zhiheng Chen

Zongci Huang

Xu Chen

Abstract

Introduction

Data and methods

Data

Table 1.

Methods

Simulation biases of WNPAC in CMIP6 multi-model ensemble

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Origins of intermodel spread in CMIP6 models simulating WNPAC

Fig. 5.

Fig. 6.

Fig. 7.

Fig. 8.

Fig. 9.

Fig. 10.

Fig. 11.

Fig. 12.

Summary and discussion

Supplementary Information

Acknowledgements

Author contributions

Funding

Data availability

Declarations

Competing interests

Footnotes

Contributor Information

Supplementary Information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases