The Winter North Pacific Teleconnection in Response to ENSO and the MJO in Operational Subseasonal Forecasting Models Is Too Weak

Chaim I. Garfinkel aFredy and Nadine Herrmann Institute of Earth Sciences, Hebrew University, Jerusalem, Israel

Search for other papers by Chaim I. Garfinkel in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0001-7258-666X
,
Wen Chen bCollege of Earth and Planetary Sciences, University of Chinese Academy of Sciences, Beijing, China
cState Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Science

Search for other papers by Wen Chen in
Current site
Google Scholar
PubMed
Close
,
Yanjie Li cState Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Science

Search for other papers by Yanjie Li in
Current site
Google Scholar
PubMed
Close
,
Chen Schwartz aFredy and Nadine Herrmann Institute of Earth Sciences, Hebrew University, Jerusalem, Israel

Search for other papers by Chen Schwartz in
Current site
Google Scholar
PubMed
Close
,
Priyanka Yadav dInstitute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland

Search for other papers by Priyanka Yadav in
Current site
Google Scholar
PubMed
Close
, and
Daniela Domeisen dInstitute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
eUniversity of Lausanne, Lausanne, Switzerland

Search for other papers by Daniela Domeisen in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0002-1463-929X
Free access

Abstract

Teleconnection patterns associated with the Madden–Julian oscillation (MJO) and El Niño–Southern Oscillation (ENSO) impact weather and climate phenomena in the Pacific–North American region and beyond, and therefore accurately simulating these teleconnections is of importance for seasonal and subseasonal forecasts. Systematic biases in boreal midwinter ENSO and MJO teleconnections are found in eight subseasonal to seasonal (S2S) forecast models over the Pacific–North America region. All models simulate an anomalous 500-hPa geopotential height response that is too weak. This overly weak response is associated with overly weak subtropical upper-level convergence and a too-weak Rossby wave source in most models, and in several models there is also a biased subtropical Pacific jet, which affects the propagation of Rossby waves. In addition to this overly weak response, all models also simulate ENSO teleconnections that reach too far poleward toward Alaska and northeastern Russia. The net effect is that these models likely underestimate the impacts associated with the MJO and ENSO over western North America, and suffer from a reduction in skill from what could be achieved.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Publisher's Note: This article was revised on 5 January 2023 to include coauthor Domeisen's ORCID number.

Corresponding author: Chaim I. Garfinkel, chaim.garfinkel@mail.huji.ac.il

Abstract

Teleconnection patterns associated with the Madden–Julian oscillation (MJO) and El Niño–Southern Oscillation (ENSO) impact weather and climate phenomena in the Pacific–North American region and beyond, and therefore accurately simulating these teleconnections is of importance for seasonal and subseasonal forecasts. Systematic biases in boreal midwinter ENSO and MJO teleconnections are found in eight subseasonal to seasonal (S2S) forecast models over the Pacific–North America region. All models simulate an anomalous 500-hPa geopotential height response that is too weak. This overly weak response is associated with overly weak subtropical upper-level convergence and a too-weak Rossby wave source in most models, and in several models there is also a biased subtropical Pacific jet, which affects the propagation of Rossby waves. In addition to this overly weak response, all models also simulate ENSO teleconnections that reach too far poleward toward Alaska and northeastern Russia. The net effect is that these models likely underestimate the impacts associated with the MJO and ENSO over western North America, and suffer from a reduction in skill from what could be achieved.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Publisher's Note: This article was revised on 5 January 2023 to include coauthor Domeisen's ORCID number.

Corresponding author: Chaim I. Garfinkel, chaim.garfinkel@mail.huji.ac.il

1. Introduction

El Niño–Southern Oscillation (ENSO) dominates the interannual variability of the tropical Pacific Ocean and atmosphere (Timmermann et al. 2018, and references therein). During its warm phase (El Niño), sea surface temperatures (SSTs) in the eastern and central tropical Pacific increase, the trade winds weaken, and tropical precipitation and the upwelling branch of the Walker circulation shift eastward toward the central Pacific. Opposite-signed changes occur during the cool phase (La Niña). While ENSO is driven mainly by coupled air–sea processes within the tropical Indo-Pacific, teleconnections of ENSO extend to the rest of the world via changes in the large-scale atmospheric circulation (Taschetto et al. 2020). These teleconnections to the North Pacific are invaluable for seasonal climate prediction over North America and Eurasia (Horel and Wallace 1981; Ropelewski and Halpert 1987; Zhang et al. 1996; Shukla et al. 2000; Wang et al. 2000; DeWeaver and Nigam 2004; Brönnimann 2007; Garfinkel et al. 2013; Scaife et al. 2017; Garfinkel et al. 2019b; Weinberger et al. 2019; Taschetto et al. 2020) and begin with a subtropical North Pacific ridge and deepened Aleutian low (Tim et al. 2017).

The physical underpinning of the extratropical Northern Hemisphere atmospheric circulation response to ENSO can be understood using the theoretical framework of poleward-propagating Rossby waves forced by anomalous upper-tropospheric tropical divergent outflow and subsequent subtropical convergence (Hoskins and Karoly 1981; Held and Kang 1987; Sardeshmukh and Hoskins 1988; Hoskins and Ambrizzi 1993; Trenberth et al. 1998; Garfinkel and Hartmann 2010; Scaife et al. 2017). After the wave reaches the extratropics, it interacts with the extratropical background flow that can modify its strength and direction of propagation. Wave–mean flow interactions can refract or duct the waves, energy can be extracted from the exit region of the subtropical East Asian jet, and synoptic eddies can interact with and either weaken or strengthen this wave train (Simmons et al. 1983; Branstator 1985; Hoskins and Ambrizzi 1993; Jin and Hoskins 1995; Held et al. 1989; Harnik et al. 2010). The net result of these interactions is the forced response to ENSO, and this forced response forms a bedrock for seasonal forecasting of weather and climate extremes (Goddard and Gershunov 2020).

Models need to represent these processes accurately in order to maximize the prediction skill associated with ENSO; however, biases can lead to less than the maximum possible skill being achieved. Model biases in the extratropical mean state can lead to biased teleconnections even for an identical tropical perturbation (Ting and Sardeshmukh 1993), while biases in the tropical mean state can lead to a poor representation of the ENSO SST anomalies and/or tropical precipitation response (Spencer and Slingo 2003; Kim et al. 2014; Ham and Kug 2015; Bayr et al. 2018, 2019; Ferrett et al. 2020; Liu et al. 2021).

The forced response due to ENSO is superposed on internal atmospheric variability. If the sample size of observed and modeled ENSO events is sufficiently large, then the effect of internal atmospheric noise can be minimized, revealing biases in the forced response. In reality, the limited record length of the observed response to ENSO necessarily implies that the “true” forced response to ENSO is, to some degree, unknown (Garfinkel et al. 2013; Deser et al. 2017). Statistical resampling techniques suggest that error bars on the forced response for events over the past century are of similar magnitude to the forced response itself (e.g., Fig. 3 of Deser et al. 2017). This uncertainty on the magnitude of the forced response leads to difficulties in ascertaining whether free-running climate models are biased in their extratropical response to ENSO, and thus whether impacts are misrepresented (Deser et al. 2018; Garfinkel et al. 2019b; Weinberger et al. 2019). Put another way, there is no reason to expect similarly phased random internal variability in a free-running model and in observations, and hence it is unclear to what extent the composite mean response to ENSO either in a model or in observations reflects the true response to ENSO versus the random superposition of internal atmospheric circulation anomalies on the forced ENSO response. That being said, systematic model biases do appear to be present in models participating in the Coupled Model Intercomparison Project (CMIP): the North Pacific low is too weak close to North America and too strong in the subtropical central and west Pacific in CMIP6 models (Fig. 2 of Fasullo et al. 2020), although in previous CMIP generations the dominant bias was a too weak teleconnection (Weare 2013) even as the ENSO events themselves extend too far to the west in most models and are too strong in many models (Ham and Kug 2015). An additional bias is that the ENSO teleconnection peaks in February in models, instead of in January (Fig. 1 of Chen et al. 2020), mirroring the bias in the decay of SST anomalies during ENSO events (Liu et al. 2021), though even the prescribed SST simulations considered by Chen et al. (2020) suffer from the delayed-teleconnection bias.

Tropical convection on time scales shorter than ENSO, due to the Madden–Julian oscillation (MJO), also can drive a North Pacific response via a mechanistically similar pathway (Seo and Son 2012; Lukens et al. 2017; Seo and Lee 2017). Specifically, the MJO can drive weather extremes in the extratropics by launching poleward propagating Rossby waves (Cassou 2008; L’Heureux and Higgins 2008; Lin et al. 2009; Johnson and Feldstein 2010; Yoo et al. 2012; Garfinkel et al. 2012b, 2014; Riddle et al. 2013; Henderson et al. 2017; Goss et al. 2018; Wang et al. 2020b). When anomalous MJO convection is in the west Pacific [MJO phases 6 and 7 as diagnosed by the Wheeler and Hendon (2004) index], the North Pacific low is deepened. These teleconnections are systematically too strong and far to the east in days 5–9 in CMIP models (Wang et al. 2020a,b), but there is substantial intermodel spread.

An alternative technique to assess whether models are systematically biased in their response to ENSO and the MJO is to use initialized reforecasts. Such reforecasts are initialized with observed sea surface temperatures and the atmospheric state (including synoptic variability); as they are intended to be useful for forecasting operationally, they can be compared directly to observed variability during the predicted lead time. Hence, they are initialized with, and meant to include, the “internal variability” that clouds the picture when considering biases in ENSO teleconnections in uninitialized climate models. Furthermore, it is crucial that operational forecasts represent these teleconnections reliably in order to maximize skill on subseasonal to seasonal time scales (Vitart 2017).

The Subseasonal-to-Seasonal (S2S) Prediction project (Vitart et al. 2017) has recently made available a large number of hindcasts covering the past several decades. This work uses these hindcasts to revisit the development of biases in ENSO teleconnections in early winter to midwinter, and the implications of these biases for MJO teleconnections. After describing the data in section 2, the biases are presented in section 3. Possible causes of these biases are also presented in section 3, and implications for the MJO are presented in section 4.

2. Data and methods

a. Data sources

The fidelity of tropics-to-extratropics teleconnections is examined in models that have contributed to the S2S Prediction project (Vitart et al. 2017). Of the 11 models that uploaded data when this study was conducted, we focus on the eight modeling centers that include output to at least week 6 to allow for model biases to develop. These are the Australian Bureau of Meteorology (BoM), the European Centre for Medium-Range Weather Forecasts (ECMWF), the China Meteorological Administration (CMA), the United Kingdom Met Office (UKMO), the National Centers for Environmental Prediction (NCEP), the Korean Meteorological Agency (KMA), the Hydrometeorological Center of Russia (HMCR), and Météo-France [Centre National de Recherches Météorologiques (CNRM)]. We downloaded hindcasts for the operational model in use during the winter of 2019/20 for all models (for ECMWF, this is CY46R1). For the ECMWF model we downloaded only one reforecast each week, and for the NCEP model we only downloaded nine reforecasts each month, for consistency with the data availability from the rest of the models. Table 1 summarizes the reforecasts analyzed in this study. ERA-Interim (ERA-I hereafter) reanalysis is used as the atmospheric reference to which forecast systems are compared (Dee et al. 2011), and NOAA v2 daily sea surface temperatures are used for the oceanic state (Huang et al. 2021).

Table 1

S2S model experiments chosen.

Table 1

We consider reforecasts initialized in December during El Niño (EN; Niño-3.4 index exceeding 0.5 K) versus those initialized in La Niña (LN; Niño-3.4 index less than −0.5 K), with the difference between these composites hereafter referred to as the response to ENSO. There are not enough ENSO events over the time period covered by the S2S archive to consider EN and LN separately (Deser et al. 2017; Garfinkel et al. 2019b), and hence we focus only on the EN minus LN difference. We focus on weekly averages as a function of week after initialization. A two-tailed difference-of-means Student’s t test is used to assess the statistical significance of the difference between EN and LN composites, and the null hypothesis is that there is no difference between the two composites. For the observed response, each year is treated as one degree of freedom. In our figures showing the modeled response, we treat each hindcast ensemble member as a separate degree of freedom as we focus on the response beyond the first week after internal variability has already led to differences among the ensemble members; accounting for the correlation among different ensemble members at these later lags has essentially no effect on the regions marked as significant (not shown).

We then consider how biases develop in the models, defined as the difference between the ENSO response in a model versus that in observations. Note that each modeling center has made available reforecasts from different years, and hence the specific ENSO events we analyze differ among the models, and further the initialization dates differ among the models even for a given year. It is therefore necessary to separately composite the observations according to the actual initializations used for each model in order to meaningfully compare the modeled and observed responses to ENSO.

The statistical significance of biases (differences between the hindcasts and ERA-I) in the response to ENSO are computed via a Monte Carlo resampling technique as follows. Taking the 500-hPa geopotential height (Z500) response to ENSO for initializations in December in ECMWF as an example, the following steps are performed.

  • Step 1: We randomly select December initializations from the full ECMWF hindcast ensemble to match the number of actual EN and LN initializations without regard to the actual ENSO phase, without allowing the same initialization to be placed in both the EN and LN fictitious samples, and without computing the ensemble mean.

  • Step 2: We then compute the difference in Z500 between the means of these fictitious EN and LN samples, both for ECMWF and for the corresponding dates in the ERA-I reanalysis.

  • Step 3: We then compute the difference in Z500 between the ECMWF and ERA-I responses for these fictitious composites.

  • Step 4: Steps 1–3 are repeated 1000 times, with different initializations randomly selected for the fictitious EN and LN composites. We then compute the bottom and top 2.5% quantiles of the difference between ECMWF and ERA-I without making any assumption as to the nature of the distribution, to which we compare the actual difference between ECMWF and ERA-I. For this resampling test, each ensemble member is treated separately, however our figures show the ensemble mean response.

The representation of the forced response to ENSO in a given model can be approximated from the response averaged over many initializations and ensemble members. Our focus in this paper is on week-6 teleconnections, which is late enough for any peculiarities of the initial condition to be mostly lost and the teleconnection to reflect model skill rather than persistence. We chose to focus on December initializations in order to understand the development of biases as the North Pacific response intensifies into midwinter, although the North Pacific response is already evident though weak in December.

A similar procedure is followed for MJO teleconnections. Specifically, the response to the MJO is computed by comparing reforecasts initialized in November–February during phases 4 and 5 and with amplitude exceeding 1 versus those initialized during phases 1 and 8 and with amplitude exceeding 1, as defined using the MJO index of Wheeler and Hendon (2004). The bias in a model is computed by comparing the MJO response in the model to that in observations. All other methodological details are as described for ENSO.

b. Diagnostics

We now describe the diagnostics used to understand biases in the response to ENSO and the MJO. The seeding of Rossby waves by tropical rainfall is diagnosed using the Rossby wave source (RWS) of Sardeshmukh and Hoskins (1988). Expressing the horizontal wind as the sum of its rotational (υψ) and divergent (υχ) components, the Rossby wave source is calculated as
RWS=(υχζ)=ζυχυχζ,
where ζ is the absolute vorticity. The RWS is a source term in the equation for absolute vorticity. Two physical processes contribute to the RWS: the change of vorticity due to vortex stretching (first term) and the advection of vorticity by the divergent part of the wind (second term). We compute these terms using daily data for each ensemble separately before averaging over ENSO phases and weeks. We focus on the RWS at 200 hPa (cf. Scaife et al. 2017; Ferrett et al. 2020). Three models suffer from severely too weak RWS and divergence at 200 hPa in their climatology (defined as the average over all available November through February hindcasts), let alone overly weak interannual variability, and the S2S archive has not made available adjacent pressure levels where the reanalysis RWS is still relatively strong. These three models are BoM, CMA, and HMCR; Figs. S1–S3 in the online supplemental material show the climatological 200-hPa RWS and divergence for these models in the first few days after initialization to demonstrate this rapid decay. Hence in order to diagnose biases in the convective response to ENSO/MJO, we also show the response of omega at 500 hPa, as all models simulate a realistic climatology for this variable in the first week.
The impact of biases in the model upper-level zonal winds on Rossby wave propagation is diagnosed by utilizing the stationary wavenumber Ks on Mercator coordinates (Karoly 1983; Hoskins and Ambrizzi 1993):
Ks=a(βMu¯M)1/2,
where the Mercator zonal wind u¯M is the time-mean 200-hPa zonal wind divided by the cosine of latitude, and a is the radius of Earth. The meridional gradient of absolute vorticity βM is defined by
βM=[2Ω1a(1cosϕϕ)2(cos2ϕu¯M)]cos2ϕa,
where ϕ is latitude and Ω is Earth’s rotational constant [following Eqs. (2.12) and (2.13) of Hoskins and Ambrizzi 1993]. Although the weekly-mean, ensemble-averaged zonal wind does not characterize the flow in a given ensemble member on a given day, previous work has found Ks to aid in understanding the behavior of stationary Rossby waves in observations and in GCMs (Seo and Lee 2017; Henderson et al. 2017; Wang et al. 2020b). On figures showing the Rossby wave source response to ENSO and biases in zonal wind, the Ks = 3 contour is overlaid, as this contour corresponds to a typical wavenumber of the anomalous convection in response to ENSO or the MJO, and the region of forbidden linear wave propagation (either because u¯M is easterly or because βM < 0) is hatched in gray.

A stationary linear Rossby wave of wavenumber k is reflected at or will decay beyond the turning latitude at which Ks = k. Furthermore, Rossby waves are refracted toward regions where Ks > k, so that localized regions where Ks is relatively large, such as the westerly jets, act as waveguides (Hoskins and Ambrizzi 1993). Such a waveguide is formed by the west Pacific jet due to its strong meridional curvature (2u/y2 ; Hoskins and Ambrizzi 1993; Henderson et al. 2017; Wang et al. 2020b). Thus Rossby waves forced by ENSO or the MJO, especially those with zonal wavenumber 4, propagate eastward along the jet and emanate at the jet exit region, while Rossby waves with zonal wavenumbers 1 and 2 can begin to propagate poleward immediately to the east of the region of forbidden propagation to the north of the jet following Ks contours 1 and 2 (Seo and Lee 2017), and thereby reach Alaska.

Regions of wave propagation can be diagnosed more quantitatively using wave ray tracing. Wave rays represent energy dispersion trajectories as specified by the group velocity (Hoskins and Karoly 1981; Lighthill 1978, p. 318). Here, we employ the ray equations based on wave theory for a horizontal nonuniform background flow (Karoly 1983; Li and Nathan 1997; Li et al. 2015) as described in detail in section 2b of Li et al. (2019). The ray trajectory is derived by integrating Eqs. (6)–(9) of Li et al. (2019) given the background wind, the starting latitude and longitude, and the initial zonal wavenumber. The background wind is smoothed using spectral triangular truncation at wavenumber 10 to remove small-scale disturbances. The ray is terminated when the total wavenumber becomes larger than 40.

In addition to these effects concerning the location of extrema, Wang et al. (2020b) use a linear baroclinic model to demonstrate the following:

  1. A stronger jet leads to stronger teleconnections, and vice versa when the jet is weaker. This is consistent with ray theory in that the amplitude of stationary Rossby waves is proportional to the speed of the mean zonal wind along a ray (Hoskins and Karoly 1981).

  2. A southward jet shift leads to a stronger teleconnection amplitude than a northward jet shift. This is reasonable as Rossby wave amplitude should be stronger when the strong absolute vorticity gradient in the jet is placed closer to the subtropical convergent winds [Frederiksen and Webster 1988; Garfinkel and Hartmann 2010, as diagnosed by the second term of Eq. (1)].

We will use these results in our interpretation of biases in ENSO and MJO teleconnections.

3. Results

We begin with the 500-hPa geopotential height response for initializations in EN as compared to LN in forecast week 6 for each model and for the reanalysis subsampled to match the specific dates included for each model (Fig. 1). Consistent with previous work discussed in the introduction, EN leads to a North Pacific low and a subtropical ridge. However, this response is systematically underestimated by all eight models. For five of the eight models the difference is statistically significant (i.e., a null hypothesis of sampling variability can be rejected) as given by the resampling test described in section 2. The subtropical ridge is also too weak in most models, and in all models the southeastern United States trough is too strong. Systematic and statistically robust biases are also evident over Eurasia as discussed in Garfinkel et al. (2019a), although our focus in this work is on the North Pacific sector.

Fig. 1.
Fig. 1.

(a)–(i) Response to ENSO in the sixth week after initialization in the December hindcast ensemble for the S2S models. For each model, the left column is for the responses in the S2S models, the center column is for ERA-I subsampled to match the specific dates included in each composite for each model, and the right column the difference between the model and reanalysis responses (i.e., the bias). Difference-of-means between EN and LN that are statistically significant at the 95% confidence level as given by Student’s t test are denoted with dots in the left and center columns; for the right column dots indicate that there is a significant difference between the modeled and observed response to ENSO as given by the resampling test described in section 2. The black box indicates the region focused upon for Figs. 2a and 2b. The multimodel mean bias is shown in (i), and dots indicate that all models agree on the sign of the response (left and center columns) or of the bias (right column).

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

The development of the North Pacific response to ENSO is shown in Figs. 2a and 2b, which shows the difference in height between EN and LN in the boxed region of Fig. 1. The observed North Pacific low strengthens by more than a factor of 2 over the weeks covered by these hindcasts, but models struggle to represent the full magnitude of this increase in the strength of the low. Specifically, while models generally are accurate in the first several weeks, by week 6 a systematic bias is evident: the North Pacific low is too weak in all models (Fig. 1i; see also Fig. S4). The envelope created by the ensemble spread of nearly all models does not encompass the reanalysis response (Fig. S5), and only two individual ensemble members out of 82 total across all models (less than 3%) show a response in the last forecasted week as strong as the observed response. Biases in late winter are relatively smaller and not systematic across models (see section 5), which we interpret to mean that the models are able to persist the ENSO related anomalies if they are present upon initialization.

Fig. 2.
Fig. 2.

Summary of the development of biases in response to ENSO, for initializations in December. (a),(b) North Pacific low (boxed region in Fig. 1). Omega at 500 hPa in the (c),(d) tropical Pacific (red box in Fig. 3) and (e),(f) subtropical Pacific (black box in Fig. 3). (left) The S2S models and (right) ERA-I subsampled to match the specific dates included in each composite for each model. Difference-of-means between EN and LN that are statistically significant at the 95% confidence level as given by Student’s t test are denoted with dots on the left column; for the right column dots indicate that there is a significant difference between the modeled and observed response to ENSO as given by the Monte Carlo test described in section 2.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

Such a bias in the development of the North Pacific response could arise from at least four factors: biases in the sea surface temperatures, too weak variability in the North Pacific regardless of the specific ENSO state, biased tropical convection in response to ENSO, or biases in the mean state through which the anomalous wave train propagates. Figures S6–S8 show that the forecasting systems are able to maintain anomalous ENSO sea surface temperatures for the duration of their subseasonal forecasts, and thus biases in teleconnections are not due to biases in the oceanic component of the forecasting systems. Hence our focus in the rest of this paper is on the atmospheric component. Figure S9 considers the standard deviation of 500-hPa height in the boxed region of Fig. 1 for each model in week 6 for each December initialization as compared to the corresponding dates in the reanalysis data. It is evident that the models simulate a realistic amount of variability in the North Pacific—a two-tailed f test of the ratio of the variance between each model and its corresponding observational benchmark is not large enough to reject a null hypothesis of no difference at the 5% confidence level for any model—and hence the too weak teleconnection to ENSO is not a by-product of a larger problem in the North Pacific sector. The other two possible factors listed at the beginning of this paragraph (i.e., biases in the convective response or in the background state) are, on the other hand, more relevant for the too weak teleconnection, as we demonstrate in sections 3a and 3b.

a. Biases in convection

We begin by considering the role of biases in convection and convective outflow. Figure 3 shows the vertical (pressure) velocity, or omega, response at 500 hPa to ENSO (EN minus LN). Upwelling is evident over the tropical central Pacific and downwelling over the Maritime Continent and subtropical central Pacific. This subtropical central Pacific downwelling is of crucial importance for the extratropical response, as this subtropical downwelling and the associated upper-level convergence are directly associated with the subtropical ridge (Hoskins and Karoly 1981; Sardeshmukh and Hoskins 1988; Tim et al. 2017; see Fig. S10 for a more detailed discussion). There is a wide diversity in the ability of models to simulate this response. BoM and NCEP both simulate stronger tropical upwelling than observed, while UKMO and CMA simulate too-weak a response. The biases are evident in all forecast weeks for most models (Figs. 2c,d; see also Fig. S4b), and in addition to BoM and NCEP, KMA also simulates overly strong tropical upwelling. Similar biases are also evident for subtropical subsidence (Figs. 2e,f; see also Fig. S4c): KMA simulates too strong subtropical downwelling in all weeks, while BoM and NCEP simulate overly strong subsidence is most weeks. Other models tend to simulate too weak subtropical downwelling, and hence it is reasonable that their extratropical response should be too weak as well. Specifically, Jiménez-Esteve and Domeisen (2019) find that the extratropical response is linear with the magnitude of the upper-level divergence anomalies, although not with the sea surface temperature anomalies.

Fig. 3.
Fig. 3.

As in Fig. 1, but for week-6 omega at 500 hPa. Note that HMCR has not uploaded omega to the S2S archive. The red and black boxes indicate the regions focused on by Figs. 2c,d and 2e,f respectively.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

As discussed in Sardeshmukh and Hoskins (1988) and in section 2, the forcing of extratropical Rossby waves is sensitive to upper-level divergent outflow and subtropical convergence, and we now focus on the response of divergence at 200 hPa to ENSO in Fig. 4. In reanalysis divergence is evident over the tropical central Pacific and convergence in the subtropical Pacific; however, both extrema of this meridional dipole are too weak in all models, and significantly so in most models. Even models that simulate a reasonable or too-strong response of omega at 500 hPa to ENSO (KMA, BoM, and NCEP; Fig. 3) simulate too weak of an upper-level divergence response. The implication is that in these three models—KMA, BoM, and NCEP—the vertical profile of the convection, and specifically upper-tropospheric divergent outflow, is not represented accurately. The S2S archive does not include the 250- or 150-hPa levels and hence we cannot assess other possible levels, and note that the reanalysis divergent outflow at 300 hPa is weak. In contrast, biases in divergent outflow in ECMWF, CNRM, and UKMO are proportionate to biases in omega at 500 hPa, implying a more reasonable convective profile.

Fig. 4.
Fig. 4.

As in Fig. 1, but for week-6 divergence at 200 hPa. The apparent lack of data for CMA and BoM is not a bug (section 2b). A black box shows the region where rays are launched for Fig. 8.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

This divergent outflow helps seed Rossby waves (Sardeshmukh and Hoskins 1988), and thus we show in Fig. 5 the Rossby wave source defined as in Sardeshmukh and Hoskins (1988) for each model and the corresponding days in reanalysis in week 6. In reanalysis a positive RWS extrema is evident near 30°N over the western and central Pacific; however, this RWS anomaly is too weak in most models, consistent with the too-weak divergence response to ENSO.

Fig. 5.
Fig. 5.

As in Fig. 1, but for week-6 RWS at 200 hPa [defined as in Eq. (1)]. The apparent lack of data for CMA and BoM is not a bug (section 2b). The region of forbidden propagation for linear stationary Rossby waves in the climatology is shown with gray stippling, and the Ks = 3 contour is shown in magenta [cf. Eq. (2)]. A black box shows the region where rays are launched for Fig. 8.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

Figures 35 focus on week 6, but biases are similar in other weeks. We demonstrate this for week 3 biases in all models in Figs. S11–S13, and the evolution of biases in omega, divergence at 200 hPa, and Rossby wave source in KMA in Figs. S14–S16.

While all models suffer from a too weak subtropical convergence response to ENSO consistent with the too weak extratropical Rossby wave train, for some models the underestimation is not statistically robust. Further, the Rossby wave source is of reasonable strength in CNRM, and only moderately too weak in ECMWF and UKMO, yet all models nonetheless suffer from a biased teleconnection. We now consider whether biases in the mean state may also contribute to a biased teleconnection pattern.

b. Extratropical mean state biases

Biases in the extratropical response to ENSO can also be associated with a biased extratropical mean state. The biases in the time mean zonal wind at 200 hPa in week 6 for December initializations are shown in Fig. 6. Figure 6 also includes the 40 and 60 m s−1 isotachs, and the region of forbidden propagation for linear stationary Rossby waves. Similar though weaker biases are present in earlier weeks (Fig. S17). Figure 7 shows Ks as calculated by Eq. (2).

Fig. 6.
Fig. 6.

Biases in U at 200 hPa in week 6 (color contours), the 40 and 60 m s−1 isotachs for the corresponding dates in ERA-i (black contours), and the climatological region of forbidden propagation for linear stationary Rossby waves (gray stippling).

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

Fig. 7.
Fig. 7.

As in Fig. 6, but for Ks in mean state.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

We first categorize the models by the nature of their biases:

  1. NCEP, HMCR, and BoM suffer from a northward shifted and eastward extended jet, with positive wind biases on the northward flank and negative wind biases on the equatorward flank of the climatological jet. Consistent with this, the region of forbidden linear wave propagation on the poleward flank of the jet (Ks undefined) extends farther eastward for NCEP and BoM; for HMCR wind biases elsewhere overwhelm this effect.

  2. CNRM and CMA suffer from a jet that does not extend far enough eastward and that is equatorward shifted.

  3. UKMO, ECMWF, and KMA have the smallest biases among the models; however, all three suffer from a jet that does not extend far enough eastward.

NCEP and BoM suffer from an eastward extended jet, and the North Pacific low is too close to North America, with negative height anomalies near the coast and positive height anomalies near the date line (Fig. 1). The linear theory of section 2b can causally connect these biases: the eastward extended jet leads to a better defined waveguide as diagnosed by Ks (Fig. 7), as wave 5 specifically is trapped within the subtropical jet all the way to North America. An additional effect in these models is a too weak teleconnection, and the poleward jet shift in these models is expected to contribute to this too weak teleconnection, as discussed in section 2b. However, these linear diagnostics do not appear capable of explaining why the wave train for NCEP extends too far poleward as compared to reanalysis. On the other hand, linear ray tracing does explain this ability of the wave train to reach Alaska. Specifically, Fig. 8a shows that wave 3 can reach Alaska using NCEP background winds, while Fig. 8c shows that the wave train is confined to midlatitudes if reanalysis winds are used. The stationary wavenumber does not take into account the meridional wind while ray tracing does; to isolate the possible effect of the meridional wind on the ability of the ray to reach Alaska, we show in Fig. 8b the rays if zonal winds from NCEP but meridional winds from ERA-I are used. It is clear that propagation to Alaska is weaker than in Fig. 8a; however, there is still more propagation than in Fig. 8c. Note that this poleward propagation appears to occur on the poleward flank of the forbidden region of wave propagation where wave 3 is allowed to propagate (Fig. 7a), and therefore the two linear diagnostics do not contradict.

Fig. 8.
Fig. 8.

Ray tracing of the extratropical response to wave 3 launched in the boxed region of Fig. 5 using (a) NCEP simulated winds in week 5 of December initializations, (b) zonal wind from NCEP and meridional winds from corresponding dates in ERA-I, and (c) both zonal and meridional winds from corresponding dates in ERA-I. Rays are launched from the boxed region in Fig. 5.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

The jet in CNRM and CMA is equatorward shifted, and while the theoretical arguments in section 2b suggest this should be expected to lead to a stronger teleconnection, the teleconnection in Fig. 1 is actually too weak. Hence the overly weak omega at 500 hPa response for both models (Fig. 3) overwhelms any effect from the midlatitude background state. Section 3a noted that CNRM simulated a RWS of reasonable strength despite an overly weak omega, and this seemingly paradoxical combination could be reconciled by the southward shift of the jet: a given upper-level divergence can seed a stronger RWS if the jet is too far south due to a stronger climatological potential vorticity meridional gradient [as diagnosed by the second term of Eq. (1); not shown]. Such a stronger teleconnection for an equatorward shifted jet was found explicitly in both the linear baroclinic model of Wang et al. (2020b) and the shallow water model of Garfinkel and Hartmann (2010). Hence the extratropical teleconnection of CNRM would likely be even weaker if the jet bias was corrected.

Finally, UKMO, ECMWF, and KMA all suffer from relatively small jet biases. However, in all three the subtropical jet does not extend far enough eastward, and consistent with this the North Pacific low does not extend far enough eastward to midlatitude North America, but rather is too strong near Alaska (Fig. 1). The biases in the subtropical jet and in the location of the North Pacific low are linked, as the waveguide formed by the subtropical jet ends too abruptly in the central Pacific, while the forbidden region for linear wave propagation ends near the date line rather than extending farther eastward (Fig. 7). This effect is demonstrated explicitly using Rossby ray tracing in Fig. S18: wave 3 launched from the west Pacific is able to reach Alaska using the KMA climatological winds, but not using reanalysis background winds.

Overall, biases in the midlatitude basic state lead to biases in teleconnections in all models, although the details of the biases differ among models and cannot be easily generalized.

4. Implications for MJO teleconnections

Stan et al. (2022) recently considered biases in MJO teleconnections in initialized forecasts using the S2S database, and found that teleconnections to the North Pacific were too weak in weeks 2 and 3 in nearly all models. At even earlier lags this too weak bias goes away and is even reversed to a too strong bias for some models (Wang et al. 2020a; Stan et al. 2022), but in week 1 the anomalous convection is already incorporated via the initialization. We now revisit the degradation in the North Pacific teleconnection in weeks 2 and 3, and then consider the role of biased tropical convection.

Figure 9 shows the geopotential height response lagged three weeks after MJO phases 4/5 as compared to phases 8/1. We focus on week 3 and these phases in order to focus on the development of the convective response and the wave train, rather than on initializations when the convective response is already present in the west Pacific. Nevertheless, results are similar for MJO phases 5/6 as compared to phases 1/2 in week 2 (Fig. S19; Stan et al. 2022): MJO phases with enhanced convection in the west Pacific lead to a deepened North Pacific low.

Fig. 9.
Fig. 9.

As in Fig. 1, but for week 3 following NDJF initializations during MJO phases 4/5 as compared to MJO phases 8/1. The black box is intentionally left in the same location as Fig. 1.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

A trough is present in observations in the North Pacific, but this trough is underestimated by all S2S models [consistent with Stan et al. (2022)]. This overly weak teleconnection is, in turn, related to biases in the tropics. Goss and Feldstein (2018) and Ting and Held (1990) find that a deeper North Pacific low occurs about a week after enhanced convection in either the west or central tropical Pacific, and hence we focus on the week-2 response in omega at 500 hPa in Fig. 10. In reanalysis, the tropical upwelling is present over the western Pacific and downwelling over the Indian Ocean; however, almost all models miss either the amplitude or the eastward propagation phase speed of the convection (or both). In NCEP, UKMO, CNRM, and CMA, the zonal dipole in omega is both too weak and too far to the west. The amplitude is better simulated by BoM and ECMWF, but the eastward phase speed is too slow (Vitart 2017). In KMA, the amplitude and phase speed are more reasonable, but the anomalous convection appears to move off of the equator in both hemispheres.

Fig. 10.
Fig. 10.

As in Fig. 3, but for week 2 following NDJF initializations during MJO phases 4/5 as compared to MJO phases 8/1. The boxes are shifted 35° to the west as compared to Fig. 3.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

The extratropical response is more sensitive to the anomalous downwelling to the south of Japan where it seeds the Rossby wave, and this anomalous downwelling is underestimated in most models in week 2. Even in the few models that represent this subtropical downwelling well in week 2 (e.g., KMA and UKMO), this downwelling is too weak in week 3 (Fig. S20). Consistent with this, the 200-hPa divergence and Rossby wave source anomalies are too weak as well (Figs. S21 and S22). Hence the quasi-steady tropical forcing leads to an extratropical teleconnection that is too weak, or simply not present, in all models.

5. Discussion

Thus far we have focused on biases in ENSO teleconnections for December initializations, and demonstrated too weak of a teleconnection in midwinter. In contrast, Chen et al. (2020) find that most CMIP models systematically simulate too strong of an ENSO teleconnection in spring. While some of the S2S models also share this bias, others do not and the multimodel mean is realistic. As an example, Figs. 11a and 11b show the response to ENSO for initializations in February, and can be compared to the comparable Figs. 2a and 2b for December initializations. In February, ECMWF simulates too weak a response at later lags while BoM simulates too strong a response (Figs. 11a,b). In other models the differences are not statistically robust, and the multimodel mean in week 6 is indistinguishable from the reanalysis response (Fig. S23). Hence late winter biases in S2S and in CMIP5 models are not similar. One possible explanation for this difference is that part of the bias in CMIP models arises from SST biases associated with the ocean coupling (e.g., too weak of a decay of ENSO in spring; Liu et al. 2021). However, Chen et al. (2020) and Garfinkel et al. (2019b) found this bias in specified SST simulations of two distinct atmospheric models. It is notable that all S2S models except CMA capture the spring tropical convective response accurately (Figs. 11c,d), although the subtropical downwelling is stronger for most models than in observations (Figs. 11e,f). Future work should consider the nature of spring biases in AMIP simulations of additional models.

Fig. 11.
Fig. 11.

As in Fig. 2, but for initializations in February.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

Goss et al. (2018) find that the extratropical response to seasonal-mean versus time-varying convective precipitation occurs due to the same physical mechanisms, and hence the mechanism leading to a North Pacific response to the MJO or to ENSO is generic to the details of how the convective anomalies develop. We find that the extratropical teleconnection response three weeks after MJO phase 4/5 and to ENSO are similar (although with the North Pacific low and North American ridge around 20° farther eastward for ENSO; Fig. 1 vs Fig. 9), even though the tropical upwelling anomalies are qualitatively different (Fig. 3 vs Fig. 10). The tropical upwelling anomalies are less important for the extratropical response than the subtropical upper-level divergence and Rossby wave source anomalies, and this subtropical response is more similar between these MJO and ENSO composites. Specifically both ENSO and the MJO feature upper-level convergence and subsidence between 20° and 30°N, although the ENSO subsidence is farther east by around 35° (Fig. 4; see also Fig. S19). The net effect is a Rossby wave source to the south of Japan two weeks after MJO phase 4/5 and over the west/central Pacific in response to ENSO (Fig. 5; see also Fig. S20). This pronounced zonal shift in the RWS is evident, but much more muted, in geopotential height, likely because linear theory prohibits a North Pacific low response over the west Pacific, and hence the Rossby wave forced by the MJO must first propagate zonally within the subtropical jet before it can effectively propagate meridionally (Seo and Lee 2017). The net effect is that the geopotential height anomalies in response to ENSO and the MJO are more similar than the underlying convective anomalies.

This paper demonstrated that biases in subtropical downwelling are related to biases in geopotential height on both intraseasonal and seasonal time scales. Biases in subtropical downwelling on even longer time scales, namely the climatology of each model, were also found to be associated with biases in climatological stationary waves in Schwartz et al. (2022). Specifically, stronger subsidence over the subtropical west Pacific is associated with a deeper low in the North Pacific between 195° and 215°E on seasonal mean time scales as well (Fig. 9 of Schwartz et al. 2022).

These results have implications for the stratospheric response to ENSO. The stratosphere is particularly sensitive to height variability in the subpolar northwest Pacific, as troughs in this region can constructively interfere with the climatological stationary waves and reinforce upward wave propagation into the stratosphere (Garfinkel et al. 2010; Domeisen et al. 2019). Figures 12a and 12b focus on geopotential height anomalies in this region in response to EN as compared to LN, and demonstrate that in the models the North Pacific low extends to this region (Fig. 12a), but in reanalysis data a ridge occurs in this region at later lags (Fig. 12b). This effect is also evident in Fig. 1: the North Pacific low extends too far northwestward toward the Russian coast in most models (see also Fig. S24). Consistent with this bias in height in the subpolar North Pacific, the stratospheric polar vortex is weakened in EN relative to LN in five models (ECMWF, UKMO, KMA, HMCR, and NCEP), but not in reanalysis data over the corresponding period. This divergent polar vortex response is demonstrated in Figs. 12c and 12d: 100-hPa polar cap height is higher for EN at 100 hPa in the models (Fig. 12c), but not for the corresponding period in reanalysis data (Fig. 12d). A similar bias in the northwestern extent of the North Pacific low was also evident for CCMVal2 models in Garfinkel et al. (2012a). Consistent with this bias in the North Pacific low, the S2S and CCMVal2 models simulate too many SSWs in EN relative to LN (Garfinkel et al. 2012a, 2019a).

Fig. 12.
Fig. 12.

As in Fig. 2, but for (a),(b) the northwest Pacific SSW precursor region, defined as 52°–72°N, 155°–185°E following Garfinkel et al. (2010). (c),(d) The 100-hPa polar cap height response.

Citation: Journal of Climate 35, 24; 10.1175/JCLI-D-22-0179.1

Vitart (2017) found that MJO teleconnections extended too far to the northwestern Pacific 11–15 days (third pentad) after a strong MJO phase 3, and to a lesser extent also 11–15 days after a strong MJO phase 7. At the lags considered in this work, this bias is no longer present, and instead the MJO teleconnection is too weak (Fig. 9).

6. Summary

ENSO and the MJO dominate the interannual and intraseasonal variability of the tropical Indo-Pacific atmosphere (Madden and Julian 1994; Timmermann et al. 2018; Jiang et al. 2020, and references therein). Teleconnections of ENSO and the MJO extend to the rest of the world (Taschetto et al. 2020) and form a basis for seasonal climate prediction over North America and East Asia among other regions (Horel and Wallace 1981; Ropelewski and Halpert 1987; Zhang et al. 1996; Shukla et al. 2000; Wang et al. 2000; DeWeaver and Nigam 2004; Brönnimann 2007; Garfinkel et al. 2013; Scaife et al. 2017; Garfinkel et al. 2019b; Domeisen et al. 2019; Weinberger et al. 2019). In order for models to actualize the potential predictability associated with ENSO and the MJO, they need to simulate the poleward propagating wave train in the North Pacific, which is seeded by convective outflow in the tropics. Hence, models need to represent the upper-level tropical divergent outflow and subtropical convergence accurately in order to maximize skill. It can be difficult to quantify the robustness of model biases in free-running climate simulations due to unrelated variability and to model biases in other processes. However, initialized forecasts offer the opportunity to study model biases in teleconnections in which many sources of unrelated variability are relatively minimized.

This study tracks the development of North Pacific teleconnections in forecasting models as compared to observations. In all models examined, the North Pacific low in response to ENSO is too weak in the North Pacific, and in most models significantly so. Less than 3% of individual ensemble members simulate a response as strong as that observed. To understand why models struggle with the magnitude of the North Pacific response, we analyzed three diagnostics of tropical outflow and subtropical convergence: the pressure velocity (omega) at 500 hPa, divergence at 200 hPa, and the Rossby wave source at 200 hPa. For most models, there is a clear connection between biases in these three metrics of the tropical response to ENSO/MJO and the subsequent extratropical wave train. Namely, after the effects of the initialization wear off (a week for the propagating MJO, and a month for the stationary ENSO) the subtropical upper-level convergence response to ENSO and the MJO is too weak in all models, leading to too weak of an extratropical wave train in the North Pacific. In most models, the tropical upwelling response is too weak as well. For ENSO, this effect is apparent when examining initializations in December before the wave train is well established. Future work should consider whether these models can represent the effect of ENSO on the propagation and intensity of the MJO found in observations (Wang and Li 2021).

Many models also simulate a North Pacific teleconnection that reaches too far poleward. Such a bias is likely related to biases in the extratropical background state: if the subtropical Pacific jet does not extend far enough eastward (e.g., UKMO, ECMWF, KMA, CNRM, and CMA), the extratropical wave train is not confined to the waveguide of the jet and spreads poleward. In a few models the subtropical jet is too far equatorward (CNRM, CMA), which in isolation would lead to too strong a teleconnection. However this tendency is compensated by too weak of a tropical divergent outflow at 200 hPa, and even for these models the teleconnection is too weak.

Despite these biases, the teleconnection is represented in a qualitatively realistic manner by nearly all models. Hence these models are likely already gaining skill from their representation of ENSO and the MJO. However there is a gap between any currently realized skill and the potential skill available, and thus improving the representation of these teleconnections is a pathway for improved model performance. Future work should consider the role of this gap in skill for the signal-to-noise paradox that is apparent in a wide range of subseasonal, seasonal, decadal, and centennial forecasting and projection models (Eade et al. 2014; Scaife and Smith 2018; Baker et al. 2018). While this paradox is commonly associated with the North Atlantic sector, it appears in the North Pacific too (Baker et al. 2018; Scaife and Smith 2018) and has been associated with models being undersensitive to external forcings and boundary conditions as found here.

Acknowledgments.

CIG and WC are supported by the ISF-NSFC joint research program (ISF Grant 3259/19 and National Natural Science Foundation of China Grant 41961144025). YJL was supported by the National Natural Science Foundation of China Grant 42175080. Support from the Swiss National Science Foundation through project PP00P2_198896 to PY and DD is gratefully acknowledged. S2S is a joint initiative of the World Weather Research Programme (WWRP) and the World Climate Research Programme (WCRP). Correspondence should be addressed to CIG (email: chaim.garfinkel@mail.huji.ac.il). We thank the reviewers for their helpful comments.

Data availability statement.

The original S2S database is hosted at ECMWF as an extension of the TIGGE database, and can be downloaded from the ECMWF server at http://apps.ecmwf.int/datasets/data/s2s/levtype=sfc/type=cf/. NOAA high-resolution SST data are provided by the NOAA/OAR/ESRL PSL, Boulder, Colorado, USA, from their Web site at https://psl.noaa.gov/data/gridded/data.noaa.oisst.v2.highres. html#detail.

REFERENCES

  • Baker, L. H., L. C. Shaffrey, R. T. Sutton, A. Weisheimer, and A. A. Scaife, 2018: An intercomparison of skill and overconfidence/underconfidence of the wintertime North Atlantic oscillation in multimodel seasonal forecasts. Geophys. Res. Lett., 45, 78087817, https://doi.org/10.1029/2018GL078838.

    • Search Google Scholar
    • Export Citation
  • Bayr, T., M. Latif, D. Dommenget, C. Wengel, J. Harlaß, and W. Park, 2018: Mean-state dependence of ENSO atmospheric feedbacks in climate models. Climate Dyn., 50, 31713194, https://doi.org/10.1007/s00382-017-3799-2.

    • Search Google Scholar
    • Export Citation
  • Bayr, T., D. I. Domeisen, and C. Wengel, 2019: The effect of the equatorial Pacific cold SST bias on simulated ENSO teleconnections to the North Pacific and California. Climate Dyn., 53, 37713789, https://doi.org/10.1007/s00382-019-04746-9.

    • Search Google Scholar
    • Export Citation
  • Branstator, G., 1985: Analysis of general circulation model sea surface temperature anomaly simulations using a linear model. Part I: Forced solutions. J. Atmos. Sci., 42, 22422254, https://doi.org/10.1175/1520-0469(1985)042<2242:AOGCMS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Brönnimann, S., 2007: The impact of El Niño/Southern Oscillation on European climate. Rev. Geophys., 45, RG3003, https://doi.org/10.1029/2006RG000199.

    • Search Google Scholar
    • Export Citation
  • Cassou, C., 2008: Intraseasonal interaction between the Madden–Julian Oscillation and the North Atlantic Oscillation. Nature, 455, 523527, https://doi.org/10.1038/nature07286.

    • Search Google Scholar
    • Export Citation
  • Chen, R., I. R. Simpson, C. Deser, and B. Wang, 2020: Model biases in the simulation of the springtime North Pacific ENSO teleconnection. J. Climate, 33, 998510 002, https://doi.org/10.1175/JCLI-D-19-1004.1.

    • Search Google Scholar
    • Export Citation
  • Dee, D., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • 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, 50595082, https://doi.org/10.1175/JCLI-D-16-0844.1.

    • Search Google Scholar
    • Export Citation
  • Deser, C., I. R. Simpson, A. S. Phillips, and K. A. McKinnon, 2018: How well do we know ENSO’s climate impacts over North America, and how do we evaluate models accordingly? J. Climate, 31, 49915014, https://doi.org/10.1175/JCLI-D-17-0783.1.

    • Search Google Scholar
    • Export Citation
  • DeWeaver, E., and S. Nigam, 2004: On the forcing of ENSO teleconnections by anomalous heating and cooling. J. Climate, 17, 32253235, https://doi.org/10.1175/1520-0442(2004)017<3225:OTFOET>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Domeisen, D. I., C. I. Garfinkel, and A. H. Butler, 2019: The teleconnection of El Niño Southern Oscillation to the stratosphere. Rev. Geophys., 57, 547, https://doi.org/10.1029/2018RG000596.

    • Search Google Scholar
    • Export Citation
  • Eade, R., D. Smith, A. Scaife, E. Wallace, N. Dunstone, L. Hermanson, and N. Robinson, 2014: Do seasonal-to-decadal climate predictions underestimate the predictability of the real world? Geophys. Res. Lett., 41, 56205628, https://doi.org/10.1002/2014GL061146.

    • Search Google Scholar
    • Export Citation
  • Fasullo, J. T., A. S. Phillips, and C. Deser, 2020: Evaluation of leading modes of climate variability in the CMIP archives. J. Climate, 33, 55275545, https://doi.org/10.1175/JCLI-D-19-1024.1.

    • Search Google Scholar
    • Export Citation
  • Ferrett, S., M. Collins, H.-L. Ren, B. Wu, and T. Zhou, 2020: The role of tropical mean-state biases in modeled winter Northern Hemisphere El Niño teleconnections. J. Climate, 33, 47514768, https://doi.org/10.1175/JCLI-D-19-0668.1.

    • Search Google Scholar
    • Export Citation
  • Frederiksen, J. S., and P. J. Webster, 1988: Alternative theories of atmospheric teleconnections and low-frequency fluctuations. Rev. Geophys., 26, 459494, https://doi.org/10.1029/RG026i003p00459.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., and D. L. Hartmann, 2010: The influence of the quasi-biennial oscillation on the North Pacific and El Niño teleconnections. J. Geophys. Res., 115, D20116, https://doi.org/10.1029/2010JD014181.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., D. L. Hartmann, and F. Sassi, 2010: Tropospheric precursors of anomalous Northern Hemisphere stratospheric polar vortices. J. Climate, 23, 32823299, https://doi.org/10.1175/2010JCLI3010.1.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., A. H. Butler, D. W. Waugh, and M. M. Hurwitz, 2012a: Why might SSWs occur with similar frequency in El Niño and La Niña winters? J. Geophys. Res., 117, D19106, https://doi.org/10.1029/2012JD017777.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., S. B. Feldstein, D. W. Waugh, C. Yoo, and S. Lee, 2012b: Observed connection between stratospheric sudden warmings and the Madden–Julian Oscillation. Geophys. Res. Lett., 39, L18807, https://doi.org/10.1029/2012GL053144.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., M. M. Hurwitz, D. W. Waugh, and A. H. Butler, 2013: Are the teleconnections of central Pacific and eastern Pacific El Niño distinct in boreal wintertime? Climate Dyn., 41, 18351852, https://doi.org/10.1007/s00382-012-1570-2.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., J. J. Benedict, and E. D. Maloney, 2014: Impact of the MJO on the boreal winter extratropical circulation vortex. Geophys. Res. Lett., 41, 60556062, https://doi.org/10.1002/2014GL061094.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., C. Schwartz, A. H. Butler, D. I. Domeisen, S.-W. Son, and I. P. White, 2019a: Weakening of the teleconnection from El Niño–Southern Oscillation to the Arctic stratosphere over the past few decades: What can be learned from subseasonal forecast models? J. Geophys. Res. Atmos., 124, 76837696, https://doi.org/10.1029/2018JD029961.

    • Search Google Scholar
    • Export Citation
  • Garfinkel, C. I., I. Weinberger, I. P. White, L. D. Oman, V. Aquila, and Y.-K. Lim, 2019b: The salience of nonlinearities in the boreal winter response to ENSO: North Pacific and North America. Climate Dyn., 52, 44294446, https://doi.org/10.1007/s00382-018-4386-x.

    • Search Google Scholar
    • Export Citation
  • Goddard, L., and A. Gershunov, 2020: Impact of El Niño on weather and climate extremes. El Niño Southern Oscillation in a Changing Climate, Geophys. Monogr., Vol. 253, Amer. Geophys. Union, 361–375, https://doi.org/10.1002/9781119548164.ch16.

  • Goss, M., and S. B. Feldstein, 2018: Testing the sensitivity of the extratropical response to the location, amplitude, and propagation speed of tropical convection. J. Atmos. Sci., 75, 639655, https://doi.org/10.1175/JAS-D-17-0132.1.

    • Search Google Scholar
    • Export Citation
  • Goss, M., S. Lee, S. B. Feldstein, and N. S. Diffenbaugh, 2018: Can ENSO-like convection force an ENSO-like extratropical response on subseasonal time scales? J. Climate, 31, 83398349, https://doi.org/10.1175/JCLI-D-17-0771.1.

    • Search Google Scholar
    • Export Citation
  • Ham, Y.-G., and J.-S. Kug, 2015: Improvement of ENSO simulation based on intermodel diversity. J. Climate, 28, 9981015, https://doi.org/10.1175/JCLI-D-14-00376.1.

    • Search Google Scholar
    • Export Citation
  • Harnik, N., R. Seager, N. Naik, M. Cane, and M. Ting, 2010: The role of linear wave refraction in the transient eddy–mean flow response to tropical Pacific SST anomalies. Quart. J. Roy. Meteor. Soc., 136, 21322146, https://doi.org/10.1002/qj.688.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., and I. S. Kang, 1987: Barotropic models of the extratropical response to El Niño. J. Atmos. Sci., 44, 35763586, https://doi.org/10.1175/1520-0469(1987)044<3576:BMOTER>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., S. W. Lyons, and S. Nigam, 1989: Transients and the extratropical response to El Niño. J. Atmos. Sci., 46, 163174, https://doi.org/10.1175/1520-0469(1989)046<0163:TATERT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Henderson, S. A., E. D. Maloney, and S.-W. Son, 2017: Madden–Julian oscillation Pacific teleconnections: The impact of the basic state and MJO representation in general circulation models. J. Climate, 30, 45674587, https://doi.org/10.1175/JCLI-D-16-0789.1.

    • Search Google Scholar
    • Export Citation
  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109, 813829, https://doi.org/10.1175/1520-0493(1981)109<0813:PSAPAW>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hoskins, B. J., and D. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci., 38, 11791196, https://doi.org/10.1175/1520-0469(1981)038<1179:TSLROA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hoskins, B. J., and T. Ambrizzi, 1993: Rossby wave propagation on a realistic longitudinally varying flow. J. Atmos. Sci., 50, 16611671, https://doi.org/10.1175/1520-0469(1993)050<1661:RWPOAR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Huang, B., C. Liu, E. Freeman, G. Graham, T. Smith, and H.-M. Zhang, 2021: Assessment and intercomparison of NOAA daily optimum interpolation sea surface temperature (DOISST) version 2.1. J. Climate, 34, 74217441, https://doi.org/10.1175/JCLI-D-21-0001.1.

    • Search Google Scholar
    • Export Citation
  • Jiang, X., and Coauthors, 2020: Fifty years of research on the Madden–Julian oscillation: Recent progress, challenges, and perspectives. J. Geophys. Res. Atmos., 125, e2019JD030911, https://doi.org/10.1029/2019JD030911.

    • Search Google Scholar
    • Export Citation
  • Jiménez-Esteve, B., and D. I. V. Domeisen, 2019: Nonlinearity in the North Pacific atmospheric response to a linear ENSO forcing. Geophys. Res. Lett., 46, 22712281, https://doi.org/10.1029/2018GL081226.

    • Search Google Scholar
    • Export Citation
  • Jin, F., and B. J. Hoskins, 1995: The direct response to tropical heating in a baroclinic atmosphere. J. Atmos. Sci., 52, 307319, https://doi.org/10.1175/1520-0469(1995)052<0307:TDRTTH>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Johnson, N. C., and S. B. Feldstein, 2010: The continuum of North Pacific sea level pressure patterns: Intraseasonal, interannual, and interdecadal variability. J. Climate, 23, 851867, https://doi.org/10.1175/2009JCLI3099.1.

    • Search Google Scholar
    • Export Citation
  • Karoly, D. J., 1983: Rossby wave propagation in a barotropic atmosphere. Dyn. Atmos. Oceans, 7, 111125, https://doi.org/10.1016/0377-0265(83)90013-1.

    • Search Google Scholar
    • Export Citation
  • Kim, S. T., W. Cai, F.-F. Jin, and J.-Y. Yu, 2014: ENSO stability in coupled climate models and its association with mean state. Climate Dyn., 42, 33133321, https://doi.org/10.1007/s00382-013-1833-6.

    • Search Google Scholar
    • Export Citation
  • L’Heureux, M. L., and R. W. Higgins, 2008: Boreal winter links between the Madden–Julian oscillation and the Arctic Oscillation. J. Climate, 21, 30403050, https://doi.org/10.1175/2007JCLI1955.1.

    • Search Google Scholar
    • Export Citation
  • Li, L., and T. R. Nathan, 1997: Effects of low-frequency tropical forcing on intraseasonal tropical–extratropical interactions. J. Atmos. Sci., 54, 332346, https://doi.org/10.1175/1520-0469(1997)054<0332:EOLFTF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Li, Y., J. Li, F. F. Jin, and S. Zhao, 2015: Interhemispheric propagation of stationary Rossby waves in a horizontally nonuniform background flow. J. Atmos. Sci., 72, 32333256, https://doi.org/10.1175/JAS-D-14-0239.1.

    • Search Google Scholar
    • Export Citation
  • Li, Y., J. Feng, J. Li, and A. Hu, 2019: Equatorial windows and barriers for stationary Rossby wave propagation. J. Climate, 32, 61176135, https://doi.org/10.1175/JCLI-D-18-0722.1.

    • Search Google Scholar
    • Export Citation
  • Lighthill, J., 1978: Waves in Fluids. Cambridge University Press, 528 pp.

  • Lin, H., G. Brunet, and J. Derome, 2009: An observed connection between the North Atlantic Oscillation and the Madden–Julian oscillation. J. Climate, 22, 364380, https://doi.org/10.1175/2008JCLI2515.1.

    • Search Google Scholar
    • Export Citation
  • Liu, M., H.-L. Ren, R. Zhang, S. Ineson, and R. Wang, 2021: ENSO phase-locking behavior in climate models: From CMIP5 to CMIP6. Environ. Res. Commun., 3, 031004, https://doi.org/10.1088/2515-7620/abf295.

    • Search Google Scholar
    • Export Citation
  • Lukens, K. E., S. B. Feldstein, C. Yoo, and S. Lee, 2017: The dynamics of the extratropical response to Madden–Julian Oscillation convection. Quart. J. Roy. Meteor. Soc., 143, 10951106, https://doi.org/10.1002/qj.2993.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., and P. R. Julian, 1994: Observations of the 40–50-day tropical oscillation—A review. Mon. Wea. Rev., 122, 814837, https://doi.org/10.1175/1520-0493(1994)122<0814:OOTDTO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Riddle, E. E., M. B. Stoner, N. C. Johnson, M. L. L’Heureux, D. C. Collins, and S. B. Feldstein, 2013: The impact of the MJO on clusters of wintertime circulation anomalies over the North American region. Climate Dyn., 40, 17491766, https://doi.org/10.1007/s00382-012-1493-y.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115, 16061626, https://doi.org/10.1175/1520-0493(1987)115<1606:GARSPP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45, 12281251, https://doi.org/10.1175/1520-0469(1988)045<1228:TGOGRF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Scaife, A. A., and D. Smith, 2018: A signal-to-noise paradox in climate science. npj Climate Atmos. Sci., 1, 28, https://doi.org/10.1038/s41612-018-0038-4.

    • Search Google Scholar
    • Export Citation
  • Scaife, A. A., and Coauthors, 2017: Tropical rainfall, Rossby waves and regional winter climate predictions. Quart. J. Roy. Meteor. Soc., 143 (702), 111, https://doi.org/10.1002/qj.2910.

    • Search Google Scholar
    • Export Citation
  • Schwartz, C., C. I. Garfinkel, P. Yadav, W. Chen, and D. I. Domeisen, 2022: Stationary waves and upward troposphere–stratosphere coupling in S2S models. Wea. Climate Dyn., 3, 679692, https://doi.org/10.5194/wcd-3-679-2022.

    • Search Google Scholar
    • Export Citation
  • Seo, K.-H., and S.-W. Son, 2012: The global atmospheric circulation response to tropical diabatic heating associated with the Madden–Julian oscillation during northern winter. J. Atmos. Sci., 69, 7996, https://doi.org/10.1175/2011JAS3686.1.

    • Search Google Scholar
    • Export Citation
  • Seo, K.-H., and H.-J. Lee, 2017: Mechanisms for a PNA-like teleconnection pattern in response to the MJO. J. Atmos. Sci., 74, 17671781, https://doi.org/10.1175/JAS-D-16-0343.1.

    • Search Google Scholar
    • Export Citation
  • Shukla, J., and Coauthors, 2000: Dynamical seasonal prediction. Bull. Amer. Meteor. Soc., 81, 25932606, https://doi.org/10.1175/1520-0477(2000)081<2593:DSP>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Simmons, A., J. M. Wallace, and G. Branstator, 1983: Barotropic wave propagation and instability, and atmospheric teleconnection patterns. J. Atmos. Sci., 40, 13631392, https://doi.org/10.1175/1520-0469(1983)040<1363:BWPAIA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Spencer, H., and J. M. Slingo, 2003: The simulation of peak and delayed ENSO teleconnections. J. Climate, 16, 17571774, https://doi.org/10.1175/1520-0442(2003)016<1757:TSOPAD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Stan, C., and Coauthors, 2022: Advances in the prediction of MJO teleconnections in the S2S forecast systems. Bull. Amer. Meteor. Soc., 103, E1426E1447, https://doi.org/10.1175/BAMS-D-21-0130.1.

    • Search Google Scholar
    • Export Citation
  • Taschetto, A. S., C. C. Ummenhofer, M. F. Stuecker, D. Dommenget, K. Ashok, R. R. Rodrigues, and S.-W. Yeh, 2020: ENSO atmospheric teleconnections. El Niño Southern Oscillation in a Changing Climate, Geophys. Monogr., Vol. 253, Amer. Geophys. Union, 309–335, https://doi.org/10.1002/9781119548164.ch14.

  • Tim, L., B. Wang, B. Wu, T. Zhou, C.-P. Chang, and R. Zhang, 2017: Theories on formation of an anomalous anticyclone in western North Pacific during El Niño: A review. J. Meteor. Res., 31, 9871006, https://doi.org/10.1007/s13351-017-7147-6.

    • Search Google Scholar
    • Export Citation
  • Timmermann, A., and Coauthors, 2018: El Niño–Southern Oscillation complexity. Nature, 559, 535545, https://doi.org/10.1038/s41586-018-0252-6.

    • Search Google Scholar
    • Export Citation
  • Ting, M., and I. M. Held, 1990: The stationary wave response to a tropical SST anomaly in an idealized GCM. J. Atmos. Sci., 47, 25462566, https://doi.org/10.1175/1520-0469(1990)047<2546:TSWRTA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Ting, M., and P. D. Sardeshmukh, 1993: Factors determining the extratropical response to equatorial diabatic heating anomalies. J. Atmos. Sci., 50, 907918, https://doi.org/10.1175/1520-0469(1993)050<0907:FDTERT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., G. W. Branstator, D. Karoly, A. Kumar, N.-C. Lau, and C. Ropelewski, 1998: Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophys. Res., 103, 14 29114 324, https://doi.org/10.1029/97JC01444.

    • Search Google Scholar
    • Export Citation
  • Vitart, F., 2017: Madden-Julian Oscillation prediction and teleconnections in the S2S database. Quart. J. Roy. Meteor. Soc., 143, 22102220, https://doi.org/10.1002/qj.3079.

    • Search Google Scholar
    • Export Citation
  • Vitart, F., and Coauthors, 2017: The Sub-Seasonal to Seasonal Prediction (S2S) Project database. Bull. Amer. Meteor. Soc., 98, 163173, https://doi.org/10.1175/BAMS-D-16-0017.1.

    • Search Google Scholar
    • Export Citation
  • Wang, B., R. Wu, and X. Fu, 2000: Pacific–East Asian teleconnection: How does ENSO affect East Asian climate? J. Climate, 13, 15171536, https://doi.org/10.1175/1520-0442(2000)013<1517:PEATHD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wang, J., H. Kim, D. Kim, S. A. Henderson, C. Stan, and E. D. Maloney, 2020a: MJO teleconnections over the PNA region in climate models. Part I: Performance- and process-based skill metrics. J. Climate, 33, 10511067, https://doi.org/10.1175/JCLI-D-19-0253.1.

    • Search Google Scholar
    • Export Citation
  • Wang, J., H. Kim, D. Kim, S. A. Henderson, C. Stan, and E. D. Maloney, 2020b: MJO teleconnections over the PNA region in climate models. Part II: Impacts of the MJO and basic state. J. Climate, 33, 50815101, https://doi.org/10.1175/JCLI-D-19-0865.1.

    • Search Google Scholar
    • Export Citation
  • Wang, T., and T. Li, 2021: Factors controlling the diversities of MJO propagation and intensity. J. Climate, 34, 65496563, https://doi.org/10.1175/JCLI-D-20-0859.1.

    • Search Google Scholar
    • Export Citation
  • Weare, B. C., 2013: El Niño teleconnections in CMIP5 models. Climate Dyn., 41, 21652177, https://doi.org/10.1007/s00382-012-1537-3.

  • Weinberger, I., C. I. Garfinkel, I. P. White, and L. D. Oman, 2019: The salience of nonlinearities in the boreal winter response to ENSO: Arctic stratosphere and Europe. Climate Dyn., 53, 45914610, https://doi.org/10.1007/s00382-019-04805-1.

    • Search Google Scholar
    • Export Citation
  • Wheeler, M. C., and H. H. Hendon, 2004: An all-season real-time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132, 19171932, https://doi.org/10.1175/1520-0493(2004)132<1917:AARMMI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yoo, C., S. Lee, and S. B. Feldstein, 2012: Mechanisms of extratropical surface air temperature change in response to the Madden–Julian oscillation. J. Climate, 25, 57775790, https://doi.org/10.1175/JCLI-D-11-00566.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, R., A. Sumi, and M. Kimoto, 1996: Impact of El Niño on the East Asian monsoon. J. Meteor. Soc. Japan, 74, 4962, https://doi.org/10.2151/jmsj1965.74.1_49.

    • Search Google Scholar
    • Export Citation

Supplementary Materials

Save
  • Baker, L. H., L. C. Shaffrey, R. T. Sutton, A. Weisheimer, and A. A. Scaife, 2018: An intercomparison of skill and overconfidence/underconfidence of the wintertime North Atlantic oscillation in multimodel seasonal forecasts. Geophys. Res. Lett., 45, 78087817, https://doi.org/10.1029/2018GL078838.

    • Search Google Scholar
    • Export Citation
  • Bayr, T., M. Latif, D. Dommenget, C. Wengel, J. Harlaß, and W. Park, 2018: Mean-state dependence of ENSO atmospheric feedbacks in climate models. Climate Dyn., 50, 31713194, https://doi.org/10.1007/s00382-017-3799-2.

    • Search Google Scholar
    • Export Citation
  • Bayr, T., D. I. Domeisen, and C. Wengel, 2019: The effect of the equatorial Pacific cold SST bias on simulated ENSO teleconnections to the North Pacific and California. Climate Dyn., 53, 37713789, https://doi.org/10.1007/s00382-019-04746-9.

    • Search Google Scholar
    • Export Citation
  • Branstator, G., 1985: Analysis of general circulation model sea surface temperature anomaly simulations using a linear model. Part I: Forced solutions. J. Atmos. Sci., 42, 22422254, https://doi.org/10.1175/1520-0469(1985)042<2242:AOGCMS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Brönnimann, S., 2007: The impact of El Niño/Southern Oscillation on European climate. Rev. Geophys., 45, RG3003, https://doi.org/10.1029/2006RG000199.

    • Search Google Scholar
    • Export Citation
  • Cassou, C., 2008: Intraseasonal interaction between the Madden–Julian Oscillation and the North Atlantic Oscillation. Nature, 455, 523527, https://doi.org/10.1038/nature07286.

    • Search Google Scholar
    • Export Citation
  • Chen, R., I. R. Simpson, C. Deser, and B. Wang, 2020: Model biases in the simulation of the springtime North Pacific ENSO teleconnection. J. Climate, 33, 998510 002, https://doi.org/10.1175/JCLI-D-19-1004.1.

    • Search Google Scholar
    • Export Citation
  • Dee, D., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, https://doi.org/10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • 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, 50595082, https://doi.org/10.1175/JCLI-D-16-0844.1.

    • Search Google Scholar
    • Export Citation
  • Deser, C., I. R. Simpson, A. S. Phillips, and K. A. McKinnon, 2018: How well do we know ENSO’s climate impacts over North America, and how do we evaluate models accordingly? J. Climate, 31, 49915014, https://doi.org/10.1175/JCLI-D-17-0783.1.

    • Search Google Scholar
    • Export Citation
  • DeWeaver, E., and S. Nigam