1. Introduction
The downward influence of major sudden stratospheric warmings (SSWs) on the troposphere is commonly described as a strongly negative phase of the North Atlantic Oscillation (NAO) with a much weaker response over the North Pacific basin (e.g., Baldwin and Dunkerton 2001; Charlton and Polvani 2007; Kolstad et al. 2010; Sigmond et al. 2013; Hitchcock and Simpson 2014). As to why the surface response to stratospheric warmings differs between the Atlantic and Pacific, there can be several potential factors. (i) Zonal asymmetries in the stratospheric anomalies associated with the initial disruption of the vortex at the time of the SSW. For this factor, it should be noted that asymmetries in the stratosphere itself rarely persist for more than a week or two while the asymmetric surface response persists for several months (Maycock and Hitchcock 2015). (ii) The dynamical differences between the tropospheric jets over the two ocean basins. One may reasonably expect the two jets to respond in different manners to the downward influence of SSWs given the fact that the Atlantic jet is primarily eddy driven while the Pacific jet is both thermally driven and eddy driven (Lee and Kim 2003; Li and Wettstein 2012). (iii) The different eddy feedbacks between the storm tracks over the two ocean basins. The Pacific sector response to SSWs might be muted if the eddy feedbacks in the Pacific storm track are weaker than those in the Atlantic. Some indication that this may be the case arises from the suppression of the midwinter storm track in the Pacific (Nakamura 1992; Penny et al. 2010; Schemm and Schneider 2018; Yuval et al. 2018). (iv) The remainder of the tropospheric precursor itself has been suggested to modulate the tropospheric response to SSWs (Charlton and Polvani 2007; Ayarzagüena et al. 2019). In particular, the different SLP responses to two types of SSWs (split/displacement) are explained based on their different tropospheric precursors, which are still present in the troposphere after the occurrence of SSWs. (v) In addition to these atmospheric internal processes shown above, the interactions between the ocean and atmosphere also play a role in shaping the surface response to SSWs. Using daily reanalysis data, Dai and Hitchcock (2021) demonstrated that the North Pacific air–sea interactions spun up by the tropospheric precursors to SSWs are responsible for the basin asymmetry in the surface response to SSWs. In particular, the atmospheric precursors in the North Pacific troposphere, through modulating the wind-driven surface heat fluxes, give rise to a sea surface temperature (SST) dipole in the North Pacific. By reinforcing the lower tropospheric baroclinicity, the SST dipole helps to sustain the generation of baroclinic eddies, strengthening the Pacific eddy-driven jet and maintaining its mean state. This prevents the Pacific eddy-driven jet from being perturbed by the downward influence of SSWs (note that the Pacific eddy-driven jet is actually more of a merged “eddy-thermally” driven jet because in the Pacific sector the eddy-driven jet is not latitudinally well separated from the subtropical jet; Lee and Kim 2003; Li and Wettstein 2012). As a result, a highly basin-asymmetric surface response in which only the Atlantic eddy-driven jet is shifted equatorward occurs, on average, in the aftermath of SSWs.
While the dynamical mechanism proposed by Dai and Hitchcock (2021) is supported by the observational evidence, it is ultimately correlative and would be further strengthened by modeling evidence. Moreover, questions remain about whether the stratosphere–troposphere–ocean coupling associated with SSWs in the real world is properly represented in state-of-the-art climate models. Quantifying the representation of SSWs in models is important as it reveals whether dynamical subseasonal-to-seasonal (S2S) prediction systems can benefit from the combined effects of stratospheric and ocean memory.
In this study, we address these questions using controlled experiments performed with a high-top version of the Community Earth System Model (CESM). A 199-yr-long free-running preindustrial simulation produces a large number of SSWs, assuring robustness of the statistical results by reducing the sampling uncertainty. On average, SSWs in the free-running preindustrial simulation reveal a basin-asymmetric negative NAO response, consistent with the reanalysis. In addition, a nudging technique is used to artificially impose a reference sudden warming in a 195-member ensemble spun off from a control simulation. This nudging methodology provides an effective means of isolating the surface response to SSWs in a comprehensive climate model (Hitchcock and Simpson 2014; Hitchcock and Haynes 2014). Similar approaches have been applied in a number of other studies to identify stratospheric influences on the tropospheric circulation (Hitchcock and Haynes 2016; Simpson et al. 2011, 2013a,b, 2018; Zhang et al. 2018; Jiménez-Esteve and Domeisen 2020; White et al. 2020). In the present work, the nudged ensembles show a highly basin-symmetric negative Northern Annular Mode (NAM) response to SSWs, in contrast to the basin-asymmetric negative NAO response in the reanalysis and the free-running preindustrial simulation, confirming that, in the absence of SSW precursors, the SSW alone would produce a more basin-symmetric response.
Following this section, a description of the data and methods used in this study is given in section 2. The large-scale tropospheric precursors and responses to SSWs are presented in section 3. It is suggested that the basin-asymmetric negative NAO response to SSWs in the free-running preindustrial simulation is shaped by the influence of their tropospheric precursors, supporting the observational results of Dai and Hitchcock (2021). Section 4 discusses the response of surface weather and climate. It turns out that, compared to SSWs with a basin-asymmetric negative NAO response, SSWs with a basin-symmetric negative NAM response have broader surface impacts, especially within and surrounding the North Pacific. The summary and discussion are presented in section 5.
2. Data and methodology
a. Model experiments
The experiments are performed with a modified version of CESM, version 1.2 (Richter et al. 2015). This version consists of the Community Atmosphere Model, version 5 (CAM5), coupled to the Parallel Ocean Program model, version 2 (POP2), and the Community Land Model, version 4 (CLM4). Instead of the default configuration of CAM5 with 30 levels and a model top at 2 hPa, we use a 46-level configuration that extends to 0.3 hPa. The atmosphere model uses the finite-volume dynamical core at approximately 0.9° × 1.25° latitude–longitude resolution.
Three sets of experiments have been performed: (i) FREE: a free-running, 199-yr-long preindustrial simulation; (ii) CTRL: a 195-yr-long preindustrial simulation, in which the zonal-mean stratospheric state is constrained to the climatology of FREE (note that both FREE and CTRL are initialized from year 402 of the 30-level CESM1 preindustrial control simulation, which is extremely similar in terms of its model physics, with the exception of the model lid height and the gravity wave drag settings, and this does not have a substantial impact on the overall climate of the model); and (iii) NUDG: a 195-member ensemble of simulations branched off from CTRL every January, in which the zonal-mean stratospheric state is nudged to follow the evolution of a reference SSW. The reference SSW is computed from a composite of SSWs associated with a polar-night jet oscillation event (hereinafter PJO SSWs) that occur in FREE. See section 2c for details of these events and a more complete justification. The NUDG integrations are started from 1 January (initialized from atmospheric and oceanic conditions taken from the CTRL run) and ended on 30 April, with the reference SSW occurring on 5 February. Nudging is performed on zonal-mean zonal wind, meridional wind, and temperature at a 6-h relaxation time scale. The model runs freely below 64 hPa and is fully constrained above 28 hPa (Simpson et al. 2018). These nudged ensembles help us to isolate the downward influence of the stratospheric anomalies on the troposphere from the influence of any preconditioning of the troposphere that may have occurred prior to the warming. By construction, differences between the NUDG and CTRL integrations in the aftermath of the reference SSW are due to the downward influence of the reference sudden warming.
The FREE and CTRL integrations described in the previous paragraph have been described and analyzed for another purpose by Simpson et al. (2018). In this study, we will compare PJO SSWs in the FREE integration with those in the reanalysis to assess the performance of the model in representing the stratosphere–troposphere–ocean coupling associated with SSWs. Additionally, we will compare PJO SSWs from the FREE and NUDG integrations to assess the influence of tropospheric precursors on the surface response to PJO SSWs.
b. Data
We use daily data from the Japanese 55-year Reanalysis (JRA-55) dataset (Ebita et al. 2011; Kobayashi et al. 2015), which has a 1.25° × 1.25° latitude–longitude resolution and 37 pressure levels from 1000 hPa up to 1 hPa. The data analyzed cover the years 1958–2019. Monthly SST data from the National Oceanic and Atmospheric Administration (NOAA) Extended Reconstructed Sea Surface Temperature version 5 (ERSSTv5) dataset (Huang et al. 2017) over the years 1958–2019 are used. The monthly ERSSTv5 data have been linearly interpolated to daily data, assuming the monthly values are centered on the 16th of the month. The temporal interpolation has been suggested to be appropriate in its own right in previous model- and observation-based studies (Taylor et al. 2000; Nigam 2003; Reynolds et al. 2007; Giorgetta et al. 2013; Butler et al. 2017).
Daily anomalies at each grid point are obtained by subtracting the seasonal cycle on that calendar day. The seasonal cycle is defined as the first three Fourier harmonics of the daily climatology.
In this study, the fields examined include sea level pressure (SLP), SST, 850-hPa zonal wind (U850-hPa), 850-hPa temperature (T850-hPa), precipitation, 850-hPa meridional temperature gradient, and 10-day high-pass-filtered eddy heat flux [defined as υ′t′, where the primes denote high-frequency components of the 850-hPa meridional wind and air temperature computed with a 101-point Lanczos filter (Duchon 1979) with a cutoff frequency of 10 days], as well as temperature, zonal wind, and geopotential height at 100 and 10 hPa.
c. Definition of SSWs
In this study, SSWs are defined by the reversal of the zonal mean westerlies at 60°N and 10 hPa following Charlton and Polvani (2007). The first day on which the daily mean zonal mean zonal wind at 60°N and 10 hPa is easterly is defined as the central date of the warming. In the following, day 0 corresponds to the SSW central date. The precursor, onset, and aftermath phases of SSWs correspond to days [−30, −3], [−2, 2], and [3, 45] around the SSW central date, respectively.
The polar-night jet oscillation (PJO) events are defined following Hitchcock et al. (2013). This involves obtaining the first two empirical orthogonal functions (EOFs) of polar-cap averaged (70°–90°N) temperature profiles. These EOFs capture the vast majority of zonal mean variability in the high-latitude Arctic stratosphere, including that associated with the stratospheric Northern Annular Mode time series often used to highlight stratosphere–troposphere interactions. The first EOF describes a vertical dipole with a lower maximum near 1 hPa; the second EOF also describes a dipole but with a lower maximum near 10 hPa. For both EOFs, we adopt the sign convention that their positive phase corresponds to a positive lower maximum. That is, the positive EOF1 corresponds to a warm upper stratosphere and the positive EOF2 corresponds to a warm middle stratosphere [see Fig. 2a in Hitchcock et al. (2013)]. We then transform the first two principal component time series (PC1 and PC2) corresponding to these two EOFs to polar coordinates r and θ, with the amplitude defined by r2 = PC12 + PC22 and the phase defined by tanθ = PC2/PC1. On this basis, a PJO event is defined to occur when the phase θ exceeds a reference phase θc, provided the amplitude r is greater than a threshold rc. This is referred to as the central date. To define the duration of the PJO event, we consider it to begin on the first date prior to the central date when the amplitude r exceeds another threshold rm and to end on the first date following the central date when the amplitude r falls below rm. Here, a reference phase of θc = 2π/3 is used, corresponding to temperature profiles with a local maximum near 60 hPa; threshold amplitudes of rc = 2σ (two standard deviations) and rm = 1.5σ (rm < rc) are used. An SSW event is considered to be associated with a PJO event if the central date of the SSW occurs during a PJO event. PJO events are nearly always associated with SSWs, and roughly half of all SSWs are associated with PJO events; these are termed PJO SSW events. PJO SSWs are significantly more persistent than non-PJO SSWs, and are responsible for the majority of the signal in composites of NAM anomalies following SSWs at lags beyond the first few weeks (Hitchcock et al. 2013). Whether an SSW is associated with a PJO event has been shown to be a more powerful stratospheric indicator of the subsequent surface response on monthly time scales than several other indicators (Karpechko et al. 2017), including whether the SSW is a split or displacement (Maycock and Hitchcock 2015).
Because we are specifically interested in the monthly-time scale surface response following SSWs, we have constructed the reference event for NUDG from PJO SSW events in FREE. Composites following PJO SSWs exhibit a substantially stronger surface signal than composites including all SSWs, so we compare the nudged response to composites of PJO SSWs in FREE and in JRA-55.
In the 61 extended winters [November–March (NDJFM)] in JRA-55, 38 SSW events are identified (Table 1). In the 198 extended winters in FREE, 94 SSW events are identified (Table 1). The frequency of SSW events per decade is 6.0 for JRA-55 and 5.0 for the FREE integration. While the model frequency is slightly less than that in JRA-55, the simulated SSW frequency is not significantly lower than the reanalysis. This is because the SSW frequency in JRA-55 (6.0 SSWs per decade) lies just within the uncertainty range (the 5th–95th percentile range) of the simulated SSW frequency, which is 4.0–6.0 SSWs per decade estimated by performing a bootstrapping test where a 61-yr sample is constructed from a random 30-yr segment and another random 31-yr segment from the FREE simulation.
Datasets to be discussed.
In JRA-55 and FREE, roughly half of all SSWs are PJO SSWs (Tables 1 and 2). As can be seen in Table 1, there are 18 PJO SSWs in JRA-55, 43 PJO SSWs in FREE, and 195 PJO SSWs in NUDG, respectively.
List of 18 central dates of PJO SSWs in JRA-55 reanalysis during winters (NDJFM) from 1958/59 to 2018/19.
d. Statistical significance assessment
1) Statistical test of composite anomalies in JRA-55 and FREE
Statistical significance of the composite analysis based on PJO SSWs in JRA-55 and FREE is assessed with a two-tailed Monte Carlo test. One thousand sets of composites are computed around N randomly chosen “central dates” from the pool of days during all of the extended winters, where N is equal to the number of PJO SSWs involved in the composite analysis. The composite anomalies based on PJO SSWs are considered significant (at the p < 0.10 level by a two-sided test) if they lie outside of the 5th–95th percentile range derived from the randomly generated composites.
2) Statistical test of composite anomalies in NUDG
The significance of composite anomalies in NUDG (the difference between the ensemble means of NUDG and CTRL) is assessed as follows. First, we calculate differences between each NUDG and CTRL member that have the same initial files; since both NUDG and CTRL have 195 members, we will get 195 differences. Then, from this pool of differences, we randomly resample 195 differences with replacement and compute their average value. This is repeated 1000 times. The uncertainty is taken as the 5th–95th percentile range and the difference between the two ensemble means is considered significant (at the p < 0.10 level by a two-sided test) if this range does not encompass zero.
3) Statistical test of differences between two groups
The significance of differences between SSWs from two groups is assessed as follows. For example, group 1 has N1 SSWs, and group 2 has N2 SSWs (N1 < N2). We randomly subsample N1 SSWs from the N2 SSWs available in group 2, with replacement, and recalculate the composite anomalies. This is repeated 1000 times. The uncertainty is taken as the 5th–95th percentile range and the difference between group 1 and group 2 is considered significant (at the p < 0.10 level by a two-sided test) if this range does not encompass the composite anomalies based on the N1 SSWs in group 1.
In this study, we are examining statistical significance at the 90% confidence level. As mentioned in section 2c, we focus on PJO SSW events, which occur half as often as all SSWs (including both PJO and non-PJO SSWs). Our willingness to accept the 90% confidence level is motivated by the relatively limited sample of PJO SSW events available in JRA-55 and in the FREE integration, particularly when considering subsets of SSWs in the FREE integration. Besides, the results presented demonstrate significance at the 90% confidence level for a variety of variables and analyses that are consistent with dynamical reasoning.
3. Large-scale tropospheric precursors and responses to PJO SSWs
a. Tropospheric precursors
To detect tropospheric precursors to PJO SSWs, we compute composites of SLP anomalies averaged over days [−30, −3] prior to the central date of PJO SSWs (Fig. 1, left panels). For the PJO SSWs in JRA-55, the SLP precursor consists of an anticyclonic center over Eurasia and a subtropical–subpolar dipole over the North Pacific (Fig. 1a), consistent with that shown in Cohen and Jones (2011) (see their Fig. 1) and Lehtonen and Karpechko (2016) (see their Fig. 2). Both features have been tied to the subsequent development of SSWs. The anticyclonic center over Eurasia corresponds to atmospheric blockings, which can help trigger SSWs (Barriopedro and Calvo 2014). The cyclonic center of the North Pacific SLP dipole is well located to constructively interfere with the climatological pressure trough (i.e., the Aleutian low) and thus weaken the vortex by enhancing upward wave propagation into the stratosphere (Garfinkel et al. 2012). On average, a similar SLP precursor pattern is seen prior to the onset of PJO SSWs in FREE (Fig. 1c), indicating that the model is capable of realistically simulating their tropospheric precursors. By contrast and by construction, the PJO SSWs in NUDG, which are artificially imposed in the stratosphere, occur without tropospheric precursors (Fig. 1e).
(left) Composites of anomalous SLP averaged over days [−30, −3] before the onset of PJO SSWs in (a) JRA-55, (c) FREE, and (e) NUDG. (right) As at left, but for anomalous SST averaged over days [−2, 2]. Solid and dashed contours denote positive and negative anomalies, respectively. Warm and cold shadings indicate positive and negative anomalies that are statistically significant at the p < 0.10 level determined with a two-tailed Monte Carlo test, respectively.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
The SLP precursor over the North Pacific has been suggested to play an additional role (Dai and Hitchcock 2021). In addition to generating SSWs, the North Pacific SLP precursor is also associated with wind-driven surface heat flux anomalies that produce a North Pacific SST dipole with above-normal SSTs at subtropical latitudes and below-normal SSTs at subpolar latitudes. Such an SST dipole can be seen at the onset of PJO SSWs in JRA-55 (Fig. 1b). A similar SST anomaly pattern can be found at the onset of PJO SSWs in FREE (Fig. 1d), although it is more like a tripole pattern that has a third center located to the south of the SST dipole (Fig. 1d). The presence of the third center also narrows down the positive lobe of the SST dipole (Fig. 1d). In addition, the positive lobe of the SST dipole is more westward shifted in the model (Fig. 1d) than in the reanalysis (Fig. 1b). These SST differences may be explained by the SLP differences in the North Pacific prior to the SSW onset between the FREE run and the reanalysis. In particular, the center of positive SLP anomalies is more westward shifted and meridionally narrowed in the model (Fig. 1c) than in the reanalysis (Fig. 1a). Unlike PJO SSWs from JRA-55 and from FREE, at the onset of PJO SSWs in NUDG, there are no substantial SST anomalies over the North Pacific (Fig. 1f). We consider next whether this SST anomaly may be shaping the subsequent tropospheric response to PJO SSWs.
b. Tropospheric circulation responses
To obtain tropospheric circulation response to PJO SSWs, we compute composites of SLP and U850-hPa anomalies averaged over days [3, 45] after the central date of PJO SSWs (Fig. 2). Following PJO SSWs in JRA-55, the tropospheric circulation response corresponds to a negative phase of NAO with a muted response in the Pacific basin (Fig. 2a). Consistently, the equatorward shift of the midlatitude jet is only seen in the Atlantic basin (Fig. 2b). This highly basin-asymmetric response is well reproduced in FREE, including the negative NAO pattern in the SLP response (Fig. 2c) and the southward shifted jet in the Atlantic basin (Fig. 2d). The good agreement between JRA-55 and FREE indicates that the model is capable of realistically reproducing the structure of the large-scale tropospheric circulation response to PJO SSWs. Even so, it is important to note that the simulated tropospheric circulation response (Figs. 2c,d) is stronger than that in the JRA-55 reanalysis (Figs. 2a,b). The difference between the model and the reanalysis exhibits a statistically significant negative NAO pattern (Figs. S1a,b), indicating that the model significantly overestimates the large-scale tropospheric circulation response to PJO SSWs. A similar model bias can be found in CESM2 (see Fig. S1 in Ayarzagüena et al. 2020). Understanding the cause and implication of this bias warrants further investigation, but is outside the scope of the present work.
Composites of anomalous (left) SLP and (right) U850-hPa averaged over days [3, 45] following PJO SSWs in (a),(b) JRA-55, (c),(d) FREE, and (e),(f) NUDG. Solid and dashed contours denote positive and negative anomalies, respectively. Warm and cold shadings indicate positive and negative anomalies that are statistically significant at the p < 0.10 level determined with a two-tailed Monte Carlo test, respectively. The two red boxes in (e) are for later use in section 3f. In the right panels, the thick red contours represent the NDJFM climatology of U850-hPa (m s−1, contour interval 6 starting from ±3) from JRA-55, FREE, and CTRL in (b), (d) and (f), respectively.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
In contrast, in the aftermath of PJO SSWs in NUDG, a more basin-symmetric tropospheric circulation response can be found (Figs. 2e,f). This corresponds to a strongly negative phase of the NAM. That is, in addition to the negative NAO response, there is a deepened Aleutian low (Fig. 2e) and an equatorward shifted jet in the North Pacific sector (Fig. 2f). Neither Pacific feature is seen in the overall response to PJO SSWs in FREE (Figs. 2c,d).
c. Dynamical attribution of the basin-asymmetric response
Dai and Hitchcock (2021) provided observational evidence that the tropospheric precursor and the resultant air–sea interactions over the North Pacific are of primary importance for shaping the basin-asymmetric negative NAO response to SSWs in the reanalysis. By contrast, ENSO teleconnection and stratospheric conditions play a rather limited role. To strengthen this argument, one must rule out other potential causes of differences between the FREE and NUDG response. We consider here tropical Pacific SSTs and stratospheric conditions associated with PJO SSWs in FREE and NUDG. The results are shown in Fig. 3.
Composites of anomalous SST averaged over days [3, 45] following PJO SSWs in (a) FREE and (b) NUDG. The magenta rectangular box denotes the Niño-3.4 region (5°S–5°N, 120°–170°W). Solid and dashed contours denote positive and negative anomalies, respectively. Orange and blue shadings indicate positive and negative anomalies that are statistically significant at the p < 0.10 level determined with a two-tailed Monte Carlo test, respectively. Temporal evolutions of zonal-averaged zonal wind along 60°N (blue curves) and area-weighted polar cap (50°–90°N) mean temperature (red curves) at (c) 10 and (d) 100 hPa. The scales of zonal wind (m s−1) and temperature (K) are labeled on the left and right axes, respectively. Black solid vertical line indicates the PJO SSW central date. Dotted curves denote the composite mean for PJO SSWs in FREE. For PJO SSWs in NUDG, thin curves indicate each individual member and thick curves denote the ensemble mean. Composites of geopotential height (m) at 10-hPa averaged over days [3, 45] following PJO SSWs in (e) FREE and (f) NUDG. (g),(h) As in (e) and (f), but for geopotential height (m) at 100 hPa.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
First, the SST anomalies in the Niño-3.4 region (5°S–5°N, 120°–170°W) are, on average, modest for both sets of PJO SSWs (Figs. 3a,b). Therefore, the distinctive features in tropospheric circulation response to PJO SSWs in FREE and NUDG (Figs. 2c–f) cannot be attributed to the remote influence arising from the tropical Pacific Ocean related to ENSO.
Second, the strength and depth of stratospheric conditions are rather consistent between the two sets of PJO SSWs (Figs. 3c,d). At 10 hPa where the zonal mean of the model is constrained in the NUDG ensembles, the ensemble mean zonal-averaged zonal wind and area-weighted polar cap temperature in NUDG (thick curves in Fig. 3c) match well with the composite mean of PJO SSWs from FREE (dotted curves in Fig. 3c). This is not surprising because the reference SSW imposed in NUDG is derived from the composite mean of PJO SSWs from FREE. Even at 100 hPa where the model is freely running in the NUDG ensembles and thus the spread among individual members is wider than that at 10 hPa, the ensemble mean in NUDG (thick curves in Fig. 3d) is only slightly different from the composite mean in FREE (dotted curves in Fig. 3d). It is unlikely that the subtle differences in the zonal-averaged stratospheric conditions between FREE and NUDG are responsible for the substantial differences in their tropospheric responses (Figs. 2c–f).
Finally, could the zonal asymmetries in the stratosphere (Figs. 3e,g) account for the basin-asymmetric negative NAO response to PJO SSWs in FREE (Figs. 2c,d)? This seems unlikely to be the case. Despite only the zonal mean component of the PJO SSW being nudged, the zonal asymmetries in the stratospheric circulation following PJO SSWs in NUDG (Figs. 3f,h) also closely resemble those in the FREE composite (Figs. 3e,g). Since the tropospheric response in NUDG is rather basin symmetric (Figs. 2e,f), zonal asymmetries in the stratosphere itself are not enough to produce a basin-asymmetric response in the troposphere.
It can therefore be concluded that the basin-asymmetric negative NAO response to PJO SSWs in FREE, similar to that in the reanalysis, mostly likely arises from the influence of the tropospheric precursor to PJO SSWs and the resultant air–sea interactions over the North Pacific.
d. Dynamical interpretation of the basin-asymmetric response
An obvious question arises from the conclusion in section 3c. That is, how do the tropospheric precursor to PJO SSWs and the resultant air–sea interactions over the North Pacific shape the basin-asymmetric negative NAO response? This question has been addressed in Dai and Hitchcock (2021) using the reanalysis dataset. Very briefly, the air–sea interactions over the North Pacific, through reinforcing the lower-tropospheric baroclinicity, help sustain the generation of baroclinic eddies, strengthening the near-surface Pacific eddy-driven jet and maintaining its near-climatological-mean state. This prevents the Pacific jet from responding to the downward influence from the stratosphere. As a result, only the Atlantic jet shifts equatorward following SSWs, exhibiting a basin-asymmetric negative NAO response. We consider next whether this dynamical mechanism is well represented or supported by the model. For this purpose, patterns of the lower tropospheric baroclinicity, synoptic eddy activity, U850-hPa, and SLP fields over the North Pacific are examined during the aftermath of PJO SSWs from FREE. These lower tropospheric fields are of interest in view of the potential for eddy–mean flow feedbacks.
Composites of these fields are computed over days [3, 45] after the central date of each PJO SSW event in FREE and then regressed against the SSW-aftermath SST index. This involves defining the daily SST index first by projecting the SST anomalies each day in extended winter (NDJFM) onto the SST anomaly pattern at the onset of PJO SSWs (Fig. 1d) for the North Pacific sector (15°–60°N, 120°E–105°W). The projected time series is normalized and used as the daily SST index. We then average the daily SST index over days [3, 45] after the central date of each PJO SSW event to obtain the SSW-aftermath SST index. By definition, the SSW-aftermath SST index has a length of 43, with each element corresponding to a PJO SSW event in FREE. The inter-SSW regression allows us to examine how the eddy-mean flow feedbacks in the aftermath of PJO SSWs vary from event to event.
As shown in Fig. 4, following those PJO SSW events that have stronger North Pacific SST anomalies, the low-level meridional temperature gradient is enhanced to the north of 30°N and weakened to the south of it (Fig. 4a). In association, synoptic eddy activity is strengthened to the north of 30°N and weakened to the south of it (Fig. 4b). Consequently, the low-level midlatitude westerlies become stronger in the region of enhanced synoptic eddy activity and weaker in the region of weakened synoptic eddy activity (Fig. 4c). This dipolar wind anomaly corresponds to an anticyclonic SLP anomaly over the North Pacific (Fig. 4d).
Regressions of anomalous lower tropospheric fields averaged over days [3, 45] following each PJO SSW event against the SSW-aftermath SST index (defined in section 3d). (a) 850-hPa meridional temperature gradient, (b) 850-hPa synoptic eddy heat flux, (c) 850-hPa zonal wind, and (d) SLP. Solid and dashed contours denote positive and negative anomalies, respectively. Orange and blue shadings indicate positive and negative anomalies that are statistically significant at the p < 0.10 level determined with a two-tailed Student’s t test, respectively. Colored contours represent the NDJFM climatology of 850-hPa meridional temperature gradient (K per degree, contour interval 0.4 starting from ±0.2) in (a), 850-hPa eddy heat flux (K m s−1; contour interval 4 starting from ±2) in (b), U850-hPa (m s−1; contour interval 4 starting from ±2) in (c), and SLP (hPa; contour interval 5.0 starting from 1000) in (d). Red and blue contours denote positive and negative values, respectively. The climatological mean fields have been smoothed with a 7-point Gaussian-weighted moving average filter to reduce the noise over land.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
In summary, the North Pacific SST anomaly spun up by the PJO SSW precursor, through regional air–sea interactions and the resultant eddy–mean flow feedbacks, favors positive wind anomalies to the north of the climatological jet core and negative wind anomalies to the south of it (Fig. 4c). This effect is to push the Pacific jet poleward, acting against the downward influence from the stratosphere that tends to push the jet equatorward. As a result, only the Atlantic jet is shifted equatorward by the downward influence from the stratosphere, resulting in a basin-asymmetric negative NAO response to PJO SSWs.
It is important to point out that the influence of the North Pacific SST anomaly on the Pacific jet shown in Fig. 4 is somewhat different from that in JRA-55 (see Fig. 9 in Dai and Hitchcock 2021). In particular, the effect in the model is to push the Pacific jet poleward, while the effect in reanalysis is to strengthen the Pacific jet locally. This can be explained by the model bias in the North Pacific SSTs (see Figs. 1b,d). As can be seen, the North Pacific SST anomaly at the onset of PJO SSWs in JRA-55 features a dipole pattern with above-normal SSTs at subtropical latitudes and below-normal SSTs at subpolar latitudes (Fig. 1b). This SST dipole strengthens the low-level meridional temperature gradient across the storm track, favoring positive surface wind anomalies that are precisely positioned to strengthen the Pacific jet (see Fig. 9 in Dai and Hitchcock 2021). By contrast, the North Pacific SST anomaly in FREE features a tripole pattern: besides the extratropical dipole part, there is a third center with below-normal SSTs to the south of it (Fig. 1d). This SST tripole strengthens and weakens the low-level meridional temperature gradient on the poleward and equatorward side of the storm track respectively, resulting in a dipolar wind anomaly that favors a poleward shift of the Pacific jet (Fig. 4). The cause of this model bias in North Pacific SSTs warrants further investigation, but is also outside the scope of the present work.
e. Identifying the basin-symmetric response in the FREE integration
In the above sections, we have shown how the influence of the tropospheric precursor over the North Pacific acts to shape the basin-asymmetric negative NAO response to PJO SSWs in the FREE integration. However, it is important to point out that the tropospheric precursor over the North Pacific is not essential to trigger SSWs. SSWs can also arise from tropospheric blocking precursors (Martius et al. 2009; Nishii et al. 2011; Barriopedro and Calvo 2014) or stratospheric internal dynamics (Scott and Polvani 2004, 2006; Hitchcock and Haynes 2016; de la Cámara et al. 2019). For those SSWs, their tropospheric responses may correspond to a basin-symmetric negative NAM rather than a basin-asymmetric NAO, given that the North Pacific tropospheric precursor to shape the basin-asymmetric negative NAO response is absent. This has been shown to be the case for SSWs in JRA-55. In particular, from all SSWs available in JRA-55, Dai and Hitchcock (2021) identified a subset of SSWs which occur without the tropospheric precursor over the North Pacific. It turns out that these SSWs are followed on average by a basin-symmetric negative NAM response, in contrast to the basin-asymmetric negative NAO response seen in the composite average of all SSWs available in JRA-55. We assess next whether these features can be found in PJO SSWs from the FREE integration.
For this purpose, an SSW-precursor SLP index is defined by averaging the SLP anomalies within the precursor region (53°–73°N, 165°–135°W; the red box in Fig. 5a) over days [−30, −3] before the onset of each PJO SSW in FREE. The precursor region is defined similarly to Garfinkel et al. (2012) (52.5°–72.5°N, 165°E–165°W; squares in their Fig. 1) but is located to the east of theirs. This is because the precursor region in this study is defined from the SLP field while the one in Garfinkel et al. (2012) is defined from the geopotential height field at 500 hPa. By definition, the SSW-precursor SLP index has a length of 43, with each element corresponding to a PJO SSW event in FREE (a histogram is shown in Fig. 5b).
(a) As in Fig. 1c: composites of anomalous SLP averaged over days [−30, −3] before the onset of PJO SSWs in FREE. The red box indicates the SSW precursor region. (b) Histogram of the SSW-precursor SLP index (defined in section 3e). The vertical solid blue line indicates the mean value over all PJO SSWs, and the black dashed vertical line denotes the zero value. (c) Scatterplot of the SSW-precursor SLP index vs the SSW-onset SST index (defined in section 3e) for the 43 PJO SSWs in FREE. The abscissa and ordinate of each dot represents the amplitude and sign of the SSW-precursor SLP index and the SSW-onset SST index, respectively. The correlation coefficient between the two indices is −0.55, as labeled in the lower-left corner of (c).
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
To confirm the connection between the North Pacific SLP precursor and the North Pacific SST anomaly in the FREE integration, an SSW-onset SST index is defined by averaging the daily SST index (defined in section 3d) over days [−2, 2] around the central date of each PJO SSW event. Similar to the SSW-precursor SLP index defined above, the SSW-onset SST index also has a length of 43, with each element corresponding to a PJO SSW event. A scatterplot of the SSW-precursor SLP index versus the SSW-onset SST index is shown in Fig. 5c. As can be seen, the two indices are negatively correlated, with a correlation coefficient of −0.55 (significant at the p < 0.01 level determined with a two-tailed Student’s t test; labeled in the lower-left corner of Fig. 5c). This scatterplot shows that those PJO SSWs with more negative SLP precursor over the North Pacific are associated with stronger North Pacific SST anomaly at the onset, evidencing the driving effect of the North Pacific SLP precursor on the formation of the North Pacific SST anomaly.
We now proceed to classify PJO SSWs in FREE according to the magnitudes of the SSW-precursor SLP index they correspond to. On average, the value of the index is slightly less than −4 hPa (blue vertical line in Fig. 5b). We therefore chose −5 hPa as the threshold value for classification. That is, a PJO SSW event is defined to occur with the SLP precursor if the value of the SSW-precursor SLP index is less than −5 hPa; it is defined to occur without the SLP precursor pattern if the value of the index is between −5 and 5 hPa. Among the 43 PJO SSW events in FREE, there are 19 PJO SSWs occurring without the North Pacific SLP precursor (referred to as “no-precursor-SSWs”) and 21 PJO SSWs with the precursor (referred to as “with-precursor-SSWs”), respectively. The remaining three PJO SSWs are preceded by high pressure anomalies within the precursor region (the SSW-precursor SLP index is greater than 5 hPa; see Fig. 5b). A comparison between no-precursor-SSWs and with-precursor-SSWs is shown in Fig. 6.
Composites of anomalous fields for two subsets of PJO SSWs in FREE. (a)–(c) 19 no-precursor-SSWs and (d)–(f) 21 with-precursor-SSWs. (top) SLP averaged over days [−30, −3], (middle) SST averaged over days [−2, 2], and (bottom) SLP averaged over days [3, 45] relative to the central date of PJO SSWs. Solid and dashed contours denote positive and negative anomalies, respectively. In (a), (b), (d), and (e), orange and blue shading indicate positive and negative anomalies that are statistically significant at the p < 0.10 level as determined with a two-tailed Monte Carlo test, respectively. In (c) and (f), stippling indicates anomalies that are statistically significant at the p < 0.10 level as determined with a two-tailed Monte Carlo test.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
First, during the precursor phase of PJO SSWs (days [−30, −3]; defined in section 2c), the SLP anomaly in the North Pacific differs substantially (by construction) between the two subsets of PJO SSWs, with weak and insignificant anomalies preceding no-precursor-SSWs (Fig. 6a) compared to large and significant anomalies preceding with-precursor-SSWs (Fig. 6d). Second, during the onset phase of PJO SSWs (days [−2, 2]; defined in section 2c), the North Pacific SST anomaly is weak and insignificant for no-precursor-SSWs (Fig. 6b) yet strong and significant for with-precursor-SSWs (Fig. 6e). This dependence of the North Pacific SST anomaly on the North Pacific SLP precursor further evidences the driving effect of the SLP precursor on the formation of the SST anomaly. Finally, during the aftermath phase of PJO SSWs (days [3, 45]; defined in section 2c), the SLP anomalies exhibit distinct features between the two subsets. In particular, a basin-symmetric negative NAM response arises in the aftermath of the no-precursor-SSWs (Fig. 6c), in contrast to the basin-asymmetric negative NAO response to those with-precursor-SSWs (Fig. 6f).
These results suggest that the basin-asymmetric negative NAO response and the basin-symmetric negative NAM response can both occur in the aftermath of PJO SSWs from the FREE integration (Fig. 6), in spite of the fact that a basin-asymmetric negative NAO response is seen, on average, following all PJO SSWs available in FREE (Figs. 2c,d). Considering all SSWs in FREE yields the same conclusions (not shown).
f. Internal variability and sampling uncertainty
In the above section, we have shown that the subset of no-precursor-SSWs in FREE exhibit a basin-symmetric negative NAM response (Fig. 6c). However, the negative NAM response to no-precursor-SSWs in FREE looks obviously different from the negative NAM response to PJO SSWs in NUDG (Fig. 2e), particularly in the Pacific sector (see Fig. 7a for their difference). Why is this so? To address this question, there is a need to highlight the large internal variability of the climate system in the Pacific sector and the relatively limited sample of no-precursor-SSWs available in the FREE integration.
(a) Difference between composites for the 19 no-precursor-SSWs in FREE and all 195 PJO SSWs in NUDG in SLP response (subtract Fig. 6c from Fig. 2e). The two red boxes are the same as those in Fig. 2e, which indicate the two SLP response regions in the Pacific and Atlantic basins used to define the basin-symmetry index (defined in section 3f). Solid and dashed contours denote positive and negative anomalies, respectively. Warm and cold shadings indicate positive and negative anomalies that are statistically significant at the p < 0.10 level determined with a two-tailed Monte Carlo test, respectively. (b) The across-member standard deviation of the SLP response to PJO SSWs in NUDG. The SLP response is defined by averaging the SLP anomalies over days [3, 45] after the onset of each PJO SSW event. (c) Histogram of 1000 sets of basin-symmetry indices assessed from 1000 sets of bootstrap-SSWs from NUDG (defined in section 3f). The red solid vertical line indicates the average value of all 1000 indices. The uncertainty is taken as the 5th–95th percentile range (indicated by the two red dashed vertical lines). The black vertical line indicates the basin-symmetry index for no-precursor-SSWs in FREE (defined in section 3e). See Fig. 6c for the composite mean SLP response to no-precursor-SSWs in FREE.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
As can be seen from the across-member standard deviation of SLP responses to PJO SSWs in NUDG (Fig. 7b), even under a common stratospheric forcing, there is a wide spread in surface responses among the 195 ensemble members, particularly in the North Pacific sector (Fig. 7b). Therefore, although the ensemble mean SLP response to all 195 PJO SSWs in NUDG is quite symmetric between the Pacific and Atlantic sectors (Fig. 2e), this is not necessarily the case for each individual PJO SSW in NUDG or the composite mean of a small number of PJO SSWs sampled from NUDG. This motivates the following bootstrap sampling. First, 19 PJO SSWs are sampled at random, with replacement, from the 195 available PJO SSWs in NUDG. Here, 19 PJO SSWs are sampled because this is the number of no-precursor-SSWs in FREE. The 19 randomly sampled PJO SSWs are referred to as “bootstrap-SSWs.” This is repeated 1000 times, resulting in 1000 sets of bootstrap-SSWs. We then compute the mean SLP response to each set of bootstrap-SSWs, obtaining 1000 sets of mean SLP responses. Second, for each set of the mean SLP response, to quantify their symmetry between the Pacific and Atlantic sectors, a basin-symmetry index is computed as the ratio of the Pacific SLP response to the Atlantic SLP response. The Pacific SLP response and the Atlantic SLP response are computed from the area-averaged SLP anomalies within the Pacific region (40°–57°N, 177°E–151°W) and the Atlantic region (42°–54°N, 30°W–13°E), respectively (the two regions are indicated by red boxes in Figs. 2e and 7a). This results in 1000 sets of basin-symmetry indices. Finally, a histogram of the 1000 sets of indices is shown in Fig. 7c. As can be seen, while the average value of the 1000 sets of indices is around 1.0 (the red solid vertical line in Fig. 7c), there is a wide spread among them. This wide spread indicates that the SLP responses to the 1000 sets of bootstrap-SSWs have basin symmetry of very different degrees, even though they share the same stratospheric forcing. This can be explained by the large internal variability in the troposphere (Fig. 7b) and the limited sample of bootstrap-SSWs.
When it comes to the subset of no-precursor-SSWs in FREE, the magnitude of the basin-symmetry index is around 0.25 (black vertical line in Fig. 7c), lying well within the uncertainty range estimated from bootstrap-SSWs in NUDG (denoted by the two dashed red lines in Fig. 7c). It therefore can be concluded that the relatively less basin-symmetric SLP response to no-precursor-SSWs in FREE (Fig. 6c), compared to the ensemble mean SLP response to all 195 PJO SSWs in NUDG (Fig. 2e), is likely due to the large internal variability of the climate system and the relatively limited sample of no-precursor-SSWs available in FREE.
g. Identifying the basin-asymmetric response in the NUDG integration
In the above section, we have shown that there is a wide spread in surface responses among the 195 NUDG ensembles, particularly in the North Pacific sector (Fig. 7b). The large inter-SSW spread in the North Pacific sector suggests that, while the ensemble mean SLP response to all 195 PJO SSWs in NUDG has an enhanced Aleutian low in the Pacific sector (Fig. 2e), this is not necessarily the case for each individual PJO SSW in NUDG. That is, some of the PJO SSWs in NUDG may have a muted response in the Pacific sector, similar to that seen in JRA-55 and FREE (Figs. 2a,c). Such PJO SSWs, if any, are likely to be preceded by Pacific anomalies that resemble the PJO SSW precursor, given the role of the PJO SSW precursor in shaping the Pacific sector response to PJO SSWs. We assess next whether such PJO SSWs can be found in the 195 PJO SSWs available in the NUDG integration.
For this purpose, the area-averaged SLP anomaly within the North Pacific precursor region (the red box in Fig. 5a) over days [−30, −3] before the onset of each PJO SSW in NUDG is calculated. We then proceed to classify PJO SSWs in NUDG according to the magnitudes of this area-averaged SLP anomaly. In particular, a PJO SSW event is defined to occur with the precursor-like SLP anomaly if the value of the area-averaged SLP anomaly is less than −5 hPa; it is defined to occur without the precursor-like SLP anomaly if the value of the area-averaged SLP anomaly is between −5 and 5 hPa. These procedures and thresholds have been used to identify with-precursor-SSWs and no-precursor-SSWs from the FREE integration (see section 3e). Here, among the 195 PJO SSW events in NUDG, 121 PJO SSWs occur without the precursor-like SLP anomaly and 33 PJO SSWs occur with the precursor-like SLP anomaly. A comparison between the two subsets of PJO SSWs is shown in Fig. 8.
Composites of anomalous fields for two subsets of PJO SSWs in NUDG. (a)–(c) 121 PJO SSWs without the precursor-like SLP anomaly and (d)–(f) 33 PJO SSWs with the precursor-like SLP anomaly. (top) SLP averaged over days [−30, −3], (middle) SST averaged over days [−2, 2], and (bottom) SLP averaged over days [3, 45] relative to the central date of PJO SSWs. Solid and dashed contours denote positive and negative anomalies, respectively. Orange and blue shading indicates positive and negative anomalies that are statistically significant at the p < 0.10 level as determined with a two-tailed Monte Carlo test, respectively.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
By construction, during the precursor phase of PJO SSWs (days [−30, −3]), the SLP anomaly over the North Pacific differs substantially between the two subsets of PJO SSWs, with weak and insignificant anomalies in the North Pacific troposphere for PJO SSWs without the precursor-like SLP anomaly (Fig. 8a) compared to large and significant anomalies preceding PJO SSWs with the precursor-like SLP anomaly (Fig. 8d). During the onset phase of PJO SSWs (days [−2, 2]), only the subset of PJO SSWs preceded by the precursor-like SLP anomaly is associated with an enhanced meridional SST gradient in the midlatitude North Pacific (Fig. 8e), which is absent in the cases without the precursor-like SLP anomaly (Fig. 8b). In the absence of the precursor-like SLP anomaly and the dipolar SST pattern, there is a highly basin-symmetric negative NAM response during the aftermath phase of PJO SSWs (Fig. 8c), similar to the ensemble mean SLP response to all 195 PJO SSWs available in NUDG (Fig. 2e). By contrast, in the presence of the precursor-like SLP anomaly and the dipolar SST pattern, a highly basin-asymmetric NAO response arises during the aftermath phase of PJO SSWs (Fig. 8f).
These results further confirm that the basin-asymmetric negative NAO response to PJO SSWs is shaped by the influence of the precursor-like SLP anomalies and the resultant air–sea interactions over the North Pacific, although those PJO SSWs in NUDG are artificially imposed in the stratosphere, rather than triggered by the precursor-like SLP anomalies.
4. Response of surface weather and climate
We have shown in Fig. 2 that, on average, tropospheric circulation response to PJO SSWs in JRA-55 and FREE exhibits a basin-asymmetric negative NAO pattern (Figs. 2a–d). By contrast, tropospheric circulation response to PJO SSWs in NUDG corresponds to a basin-symmetric negative NAM pattern (Figs. 2e,f). These distinct large-scale tropospheric circulation responses to PJO SSWs can be expected to result in very different impacts on surface weather and climate. To examine these surface impacts, we compute composites of T850-hPa, precipitation, and SST anomalies averaged over days [3, 45] after the central date of PJO SSWs (Figs. 9 and 10).
Composites of anomalous (left) T850-hPa and (right) precipitation averaged over days[3, 45] for PJO SSWs in (a),(b) JRA-55, (c),(d) FREE, and (e),(f) NUDG. Solid and dashed contours denote positive and negative anomalies, respectively. Warm and cold shadings indicate positive and negative anomalies that are statistically significant at the p < 0.10 level determined with a two-tailed Monte Carlo test, respectively.
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
Composites of anomalous SST averaged over days [3, 45] following PJO SSWs in (a) JRA-55, (b) FREE, and (c) NUDG. Warm and cold shadings denote positive and negative anomalies, respectively. Stippling indicates anomalies that are statistically significant at the p < 0.10 level as determined with a two-tailed Monte Carlo test. The color contours in (c) indicate the regressed SST anomalies onto the PDO index (defined in section 4c; contour interval 0.2 K starting from ±0.1 K).
Citation: Journal of Climate 36, 3; 10.1175/JCLI-D-22-0300.1
a. Surface air temperature
In the aftermath of PJO SSWs in JRA-55 (Fig. 9a), the temperature anomalies consist of anomalous warming over Greenland, eastern Canada, subtropical Africa, and Asia, as well as anomalous cooling over northern Europe and the eastern United States. The model reproduces qualitatively well the temperature response surrounding the North Atlantic Ocean but largely misses the surface cooling over the eastern United States (Fig. 9c). This model bias of missing surface cooling over the eastern United States is also seen in a couple of CMIP6 models analyzed in Ayarzagüena et al. (2020), including CESM2, CNRM-ESM2-1, and GFDL-CM4 [see Fig. S1 in Ayarzagüena et al. (2020) for temperature response to PJO SSWs in CMIP6 models]. Even so, for the model used in this study, the warm bias over the eastern United States is not statistically significant (Fig. S1c).
In the aftermath of PJO SSWs in NUDG, besides the typical temperature response surrounding the North Atlantic Ocean, there is substantial surface warming over Alaska and cooling over East Asia (Fig. 9e). Both features can be explained by the strengthened Aleutian low associated with the basin-symmetric negative NAM response (Fig. 2e). The intensification of the Aleutian low could increase the advection of relatively warm and moist air to Alaska, leading to surface warming (Hartmann and Wendler 2005), and at the same time it increases the pressure gradient between the Siberian high and the Aleutian low, which could strengthen the East Asian winter monsoon (EAWM) (Wang and Chen 2014). The enhanced EAWM further favors a decrease in near-surface air temperature over East Asia and a potential increase in the likelihood of cold air outbreaks over there. A recent study by Huang et al. (2021) shows that the risk of severe cold air outbreaks in midlatitude East Asia is increased by 100% during weak stratospheric polar vortex conditions, which are associated with a basin-symmetric negative NAM response in the troposphere (see Fig. S2 in Huang et al. 2021). This basin-symmetric negative NAM response (Fig. S2 in Huang et al. 2021) also consists of a strengthened Aleutian low in the Pacific sector, similar to that seen in the tropospheric response to PJO SSWs in NUDG (Fig. 2e).
b. Precipitation
The precipitation responses also differ among the three sets of PJO SSWs (Fig. 9, right panels). Following PJO SSWs in JRA-55, the marine precipitation response is largely confined within the North Atlantic basin (Fig. 9b), which exhibits a dipolar pattern with above-normal precipitation at midlatitudes and below-normal precipitation at subpolar latitudes, consistent with the southward shift of the North Atlantic jet (Fig. 2b). The terrestrial precipitation response consists of anomalously wet conditions over western and central Europe and dry conditions over Scandinavia and large regions in Asia (Fig. 9b). The model reproduces most of the precipitation responses in JRA-55 but shows extra impacts (Fig. 9d), including anomalously dry conditions over the southern United States, a basinwide dipolar response over the Mediterranean, and an elongated band of above-normal precipitation over the western North Pacific. The increased precipitation over the western North Pacific is likely a response to the variability of the underlying ocean because it corresponds with local maxima of SST warming in the western North Pacific (Fig. 10b). In the aftermath of PJO SSWs in NUDG, there is an elongated band of above-normal precipitation over the central and eastern North Pacific (Fig. 9f), consistent with the southward shift of the North Pacific jet (Fig. 2f).
To sum up, in the aftermath of PJO SSWs in JRA-55 and FREE, due to the presence of the basin-asymmetric negative NAO response, changes in surface temperature and precipitation are largely Atlantic-focused (Figs. 9a–d). By contrast, in the aftermath of PJO SSWs in NUDG, the basin-symmetric negative NAM response leads to changes in surface temperature and precipitation in both the Atlantic and the Pacific basins (Figs. 9e,f).
c. Sea surface temperature
Owing to their extraordinary persistence on time scales of weeks to months, the tropospheric circulation response to PJO SSWs is also capable of perturbing the underlying oceans.
In the North Atlantic Ocean, an SST tripole is seen following PJO SSWs in JRA-55 (Fig. 10a). The SST tripole spans the North Atlantic basin from subpolar to subtropical latitudes, consisting of below-normal SSTs between 30° and 45°N surrounded by above-normal SSTs between 15°–30°N and 45°–60°N. This SST tripole closely resembles that associated with a negative NAO pattern (Seager et al. 2000; Sutton et al. 2001; Czaja et al. 2003), consistent with the fact that PJO SSWs in JRA-55 corresponds to a negative NAO response (Fig. 2a). Due to the fact that a negative NAO pattern is also seen in the aftermath of PJO SSWs in FREE and NUDG (Figs. 2c,e), a similar SST tripole in the North Atlantic Ocean is seen in simulations (Figs. 10b,c), although the simulated SST tripole is stronger along the coast of Europe and Africa (Figs. 10b,c), in contrast to the observed SST tripole, which is stronger along the coast of America (Fig. 10a).
In the North Pacific Ocean, there are substantial SST anomalies in the aftermath of PJO SSWs in JRA-55 and FREE (Figs. 10a,b). These North Pacific SST anomalies have been suggested to be driven by the tropospheric precursor to PJO SSWs, rather than caused by their subsequent downward influence (Dai and Hitchcock 2021). The reason is twofold: 1) the North Pacific SST anomalies have been in existence at the onset of PJO SSWs (Figs. 1b,d), before their knock-on effects on the tropospheric circulation start to emerge; and 2) the tropospheric circulation response to PJO SSWs in JRA-55 and FREE corresponds to a basin-asymmetric negative NAO pattern (Figs. 2a–d), which is confined to the North Atlantic basin and therefore unlikely to perturb the North Pacific Ocean.
In the aftermath of PJO SSWs in NUDG, there is a horseshoe-shaped pattern of SST anomaly in the North Pacific (Fig. 10c, black contours and color shadings), which consists of below-normal SSTs in the central North Pacific surrounded by above-normal SSTs off the coast of the western United States and over the eastern subtropical Pacific. Unlike the North Pacific SST anomalies in JRA-55 and FREE, the horseshoe-shaped SST anomaly in NUDG does not show up at the onset of PJO SSWs (Fig. 1f), indicating that it is more likely a response to the downward influence of PJO SSWs. Furthermore, the horseshoe-shaped SST anomaly is very reminiscent of the SST pattern associated with the positive Pacific decadal oscillation (PDO) phase (Kren et al. 2016). A direct comparison between the positive PDO-related SST anomaly and the SSW-induced SST anomaly is shown in Fig. 10c, wherein the regressed SST anomaly onto the PDO index (Fig. 10c, color contours) is superimposed on the composite SST anomaly based on PJO SSWs in NUDG (Fig. 10c, black contours and color shadings). The regressed SST anomaly onto the PDO index is assessed from the CTRL integration (introduced in section 2a). The PDO index is defined as the normalized leading principal component of SST anomalies averaged over days [3, 45] after 5 February in the region 20°–60°N, 110°E–100°W of the North Pacific. The SST anomalies averaged over days [3, 45] after 5 February are then regressed onto the PDO index. As shown in Fig. 10c, the SSW-induced SST pattern (shading) bears a striking resemblance to the positive PDO-related SST pattern (color contour). This can be explained by the strengthened Aleutian low associated with the basin-symmetric negative NAM response to PJO SSWs in NUDG (Fig. 2e), given the fact that the positive PDO phase is also marked by a strengthened Aleutian low (Kren et al. 2016). Previous studies linking the positive PDO phase to weak stratospheric vortex events have been devoted to understanding the forcing of the PDO on the stratosphere (Kren et al. 2016; Ayarzagüena et al. 2021), the results here suggests that the stratospheric forcing has the potential to modulate the PDO as well. It is important to note that the PDO variability has a broad power spectrum, with significant variance at both subseasonal and decadal time scales (Newman et al. 2016), and the results here indicate a direct modulation of SSWs on the subseasonal time scales. It is worth investigating whether this subseasonal influence can give rise to a significant modulation of the decadal variability of North Pacific SST, given the decadal variability in the frequency of SSWs (Dimdore-Miles et al. 2021), which has been demonstrated as a source of decadal variability within the Atlantic meridional overturning circulation due to the downward influence over the Atlantic Ocean (Reichler et al. 2012).
5. Conclusions
In this study, a high-top version of CESM1.2 is used to investigate the downward influence of PJO SSWs with a particular focus on the dynamics and impacts of the North Pacific eddy-driven jet response. A free-running, 199-yr-long preindustrial simulation (FREE) has been performed, in which a total of 43 PJO SSWs occurred. In addition, a 195-member nudged ensemble (NUDG), spun off from a long control simulation, is performed, of which each individual member has a reference PJO SSW artificially imposed in the stratosphere. The reference PJO SSW is computed from the composite mean of PJO SSWs in FREE.
On average, the PJO SSWs in FREE are preceded by tropospheric precursors similar to that in the JRA-55 reanalysis and are followed by a basin-asymmetric negative NAO response. By contrast, the PJO SSWs in NUDG occur without tropospheric precursors and correspond to a basin-symmetric negative NAM response, which consists of a negative NAO pattern and a strengthened Aleutian low.
The basin-symmetric negative NAM response to PJO SSWs in NUDG suggests that in the absence of tropospheric precursors, the downward influence on the troposphere from the sudden warming is highly symmetric between the Pacific and Atlantic basins. However, in the JRA-55 reanalysis and the FREE integration, the overall tropospheric response to PJO SSWs is highly basin-asymmetric, which consists of a negative NAO pattern and a muted response in the Pacific basin. This highly basin-asymmetric surface response is shaped by the influence of the tropospheric precursor to PJO SSWs which, through air–sea interactions over the North Pacific, opposes the downward influence of the PJO SSW on the Pacific jet. In fact, in both JRA-55 and FREE, only those PJO SSWs preceded by North Pacific tropospheric precursors correspond to a basin-asymmetric negative NAO response; for those PJO SSWs occurring without North Pacific tropospheric precursors, a basin-symmetric negative NAM response can be expected.
Because the negative NAO response and the negative NAM response can both occur in the reanalysis and in the simulations, forecasting which response is more likely to emerge for an ongoing or upcoming PJO SSW event is crucially important. This may be achieved through monitoring atmospheric circulation and ocean temperature in the North Pacific both before and at the onset of the PJO SSW, given that the tropospheric precursor to PJO SSWs and the resultant air-sea interactions over the North Pacific are of primary importance for shaping the basin-asymmetric NAO response to PJO SSWs. This can be further used to improve predictions of surface temperatures and hydrology on S2S time scales. This is because, following PJO SSWs with a basin-asymmetric negative NAO response, changes in surface temperature and precipitation are mainly Atlantic-focused. By contrast, following PJO SSWs with a basin-symmetric negative NAM response, changes in temperature and precipitation can be found in both the Atlantic and Pacific basins. In this sense, most current operational S2S prediction systems, properly initialized from the observed atmosphere and ocean state, have the potential to forecast whether a basin-asymmetric negative NAO or a basin-symmetric negative NAM response is more likely to emerge for an ongoing or upcoming PJO SSW event. Even so, the full realization of this skill in initialized S2S predictions may be hampered by model biases in capturing the evolution of the stratospheric polar vortices and their coupling to the troposphere. Examining whether current S2S systems do indeed capture this mechanism would be a valuable next step.
On longer time scales, PJO SSWs are linked to decadal variations in the North Pacific. In particular, PJO SSWs with a basin-symmetric negative NAM response, through strengthening the Aleutian low, give rise to a positive PDO-like SST pattern in the North Pacific; this suggests a possible stratospheric forcing on the PDO.
Under higher CO2 forcing, the average surface response to SSWs will be more like a basin-symmetric negative NAM than a basin-asymmetric negative NAO (Ayarzagüena et al. 2020), indicating that SSWs with a basin-symmetric negative NAM response may happen more frequently under a warmer climate. The dynamical mechanisms provided in this study may be of relevance for understanding this change.
Acknowledgments.
Support from Cornell University is gratefully acknowledged. IRS is supported by the National Center for Atmospheric Research, which is a major facility sponsored by the National Science Foundation under the Cooperative Agreement 1852977. We would also like to acknowledge high-performance computing support provided by NCAR’s Computational and Information Systems Laboratory.
Data availability statement.
The JRA-55 reanalysis data used in this study were obtained from https://rda.ucar.edu/datasets/ds628.0/ and the NOAA ERSST.v5 data from https://www.esrl.noaa.gov/psd/data/gridded/data.noaa.ersst.v5.html. The CESM data are available from the corresponding author upon request.
REFERENCES
Ayarzagüena, B., F. M. Palmeiro, D. Barriopedro, N. Calvo, U. Langematz, and K. Shibata, 2019: On the representation of major stratospheric warmings in reanalyses. Atmos. Chem. Phys., 19, 9469–9484, https://doi.org/10.5194/acp-19-9469-2019.
Ayarzagüena, B., and Coauthors, 2020: Uncertainty in the response of sudden stratospheric warmings and stratosphere–troposphere coupling to quadrupled CO2 concentrations in CMIP6 models. J. Geophys. Res. Atmos., 125, e2019JD032345, https://doi.org/10.1029/2019JD032345.
Ayarzagüena, B., E. Manzini, N. Calvo, and D. Matei, 2021: Interaction between decadal-to-multidecadal oceanic variability and sudden stratospheric warmings. Ann. N. Y. Acad. Sci., 1504, 215–229, https://doi.org/10.1111/nyas.14663.
Baldwin, M. P., and T. J. Dunkerton, 2001: Stratospheric harbingers of anomalous weather regimes. Science, 294, 581–584, https://doi.org/10.1126/science.1063315.
Barriopedro, D., and N. Calvo, 2014: On the relationship between ENSO, stratospheric sudden warmings, and blocking. J. Climate, 27, 4704–4720, https://doi.org/10.1175/JCLI-D-13-00770.1.
Butler, A. H., J. P. Sjoberg, D. J. Seidel, and K. H. Rosenlof, 2017: A sudden stratospheric warming compendium. Earth Syst. Sci. Data, 9, 63–76, https://doi.org/10.5194/essd-9-63-2017.
Charlton, A. J., and L. M. Polvani, 2007: A new look at stratospheric sudden warmings. Part I: Climatology and modeling benchmarks. J. Climate, 20, 449–469, https://doi.org/10.1175/JCLI3996.1.
Cohen, J., and J. Jones, 2011: Tropospheric precursors and stratospheric warmings. J. Climate, 24, 6562–6572, https://doi.org/10.1175/2011JCLI4160.1.
Czaja, A., A. W. Robertson, and T. Huck, 2003: The role of the Atlantic ocean–atmosphere coupling in affecting North Atlantic Oscillation variability. The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophys. Monogr., Vol. 134, Amer. Geophys. Union, 147–172.
Dai, Y., and P. Hitchcock, 2021: Understanding the basin asymmetry in surface response to sudden stratospheric warmings from an ocean–atmosphere coupled perspective. J. Climate, 34, 8683–8698, https://doi.org/10.1175/JCLI-D-21-0314.1.
de la Cámara, A., T. Birner, and J. R. Albers, 2019: Are sudden stratospheric warmings preceded by anomalous tropospheric wave activity? J. Climate, 32, 7173–7189, https://doi.org/10.1175/JCLI-D-19-0269.1.
Dimdore-Miles, O., L. Gray, and S. Osprey, 2021: Origins of multi-decadal variability in sudden stratospheric warmings. Wea. Climate Dyn., 2, 205–231, https://doi.org/10.5194/wcd-2-205-2021.
Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18, 1016–1022, https://doi.org/10.1175/1520-0450(1979)018<1016:LFIOAT>2.0.CO;2.
Ebita, A., and Coauthors, 2011: The Japanese 55-Year Reanalysis “JRA-55”: An interim report. SOLA, 7, 149–152, https://doi.org/10.2151/sola.2011-038.
Garfinkel, C. I., A. H. Butler, D. W. Waugh, M. M. Hurwitz, and L. M. Polvani, 2012: Why might stratospheric sudden warmings occur with similar frequency in El Niño and La Niña winters? J. Geophys. Res., 117, D19106, https://doi.org/10.1029/2012JD017777.
Giorgetta, M. A., and Coauthors, 2013: Climate and carbon cycle changes from 1850 to 2100 in MPI-ESM simulations for the Coupled Model Intercomparison Project phase 5. J. Adv. Model. Earth Syst., 5, 572–597, https://doi.org/10.1002/jame.20038.
Hartmann, B., and G. Wendler, 2005: The significance of the 1976 Pacific climate shift in the climatology of Alaska. J. Climate, 18, 4824–4839, https://doi.org/10.1175/JCLI3532.1.
Hitchcock, P., and P. H. Haynes, 2014: Zonally symmetric adjustment in the presence of artificial relaxation. J. Atmos. Sci., 71, 4349–4368, https://doi.org/10.1175/JAS-D-14-0013.1.
Hitchcock, P., and I. Simpson, 2014: The downward influence of sudden stratospheric warmings. J. Atmos. Sci., 71, 3856–3876, https://doi.org/10.1175/JAS-D-14-0012.1.
Hitchcock, P., and P. H. Haynes, 2016: Stratospheric control of planetary waves. Geophys. Res. Lett., 43, 11 884–11 892, https://doi.org/10.1002/2016GL071372.
Hitchcock, P., T. G. Shepherd, and G. L. Manney, 2013: Statistical characterization of Arctic polar-night jet oscillation events. J. Climate, 26, 2096–2116, https://doi.org/10.1175/JCLI-D-12-00202.1.
Huang, B., and Coauthors, 2017: Extended Reconstructed Sea Surface Temperature version 5 (ERSSTv5): Upgrades, validations, and intercomparisons. J. Climate, 30, 8179–8205, https://doi.org/10.1175/JCLI-D-16-0836.1.
Huang, J., P. Hitchcock, A. C. Maycock, C. M. McKenna, and W. Tian, 2021: Northern Hemisphere cold air outbreaks are more likely to be severe during weak polar vortex conditions. Commun. Earth Environ., 2, 147, https://doi.org/10.1038/s43247-021-00215-6.
Jiménez-Esteve, B., and D. I. V. Domeisen, 2020: Nonlinearity in the tropospheric pathway of ENSO to the North Atlantic. Wea. Climate Dyn., 1, 225–245, https://doi.org/10.5194/wcd-1-225-2020.
Karpechko, A. Y., P. Hitchcock, D. H. W. Peters, and A. Schneidereit, 2017: Predictability of downward propagation of major sudden stratospheric warmings. Quart. J. Roy. Meteor. Soc., 143, 1459–1470, https://doi.org/10.1002/qj.3017.
Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5–48, https://doi.org/10.2151/jmsj.2015-001.
Kolstad, E. W., T. Breiteig, and A. A. Scaife, 2010: The association between stratospheric weak polar vortex events and cold air outbreaks in the Northern Hemisphere. Quart. J. Roy. Meteor. Soc., 136, 886–893, https://doi.org/10.1002/qj.620.
Kren, A., D. Marsh, A. Smith, and P. Pilewskie, 2016: Wintertime Northern Hemisphere response in the stratosphere to the Pacific decadal oscillation using the Whole Atmosphere Community Climate Model. J. Climate, 29, 1031–1049, https://doi.org/10.1175/JCLI-D-15-0176.1.
Lee, S., and H.-K. Kim, 2003: The dynamical relationship between subtropical and eddy-driven jets. J. Atmos. Sci., 60, 1490–1503, https://doi.org/10.1175/1520-0469(2003)060<1490:TDRBSA>2.0.CO;2.
Lehtonen, I., and A. Y. Karpechko, 2016: Observed and modeled tropospheric cold anomalies associated with sudden stratospheric warmings. J. Geophys. Res. Atmos., 121, 1591–1610, https://doi.org/10.1002/2015JD023860.
Li, C., and J. J. Wettstein, 2012: Thermally driven and eddy-driven jet variability in reanalysis. J. Climate, 25, 1587–1596, https://doi.org/10.1175/JCLI-D-11-00145.1.
Martius, O., L. M. Polvani, and H. C. Davies, 2009: Blocking precursors to stratospheric sudden warming events. Geophys. Res. Lett., 36, L14806, https://doi.org/10.1029/2009GL038776.
Maycock, A. C., and P. Hitchcock, 2015: Do split and displacement sudden stratospheric warmings have different annular mode signatures? Geophys. Res. Lett., 42, 10 943–10 951, https://doi.org/10.1002/2015GL066754.
Nakamura, H., 1992: Midwinter suppression of baroclinic wave activity in the Pacific. J. Atmos. Sci., 49, 1629–1642, https://doi.org/10.1175/1520-0469(1992)049<1629:MSOBWA>2.0.CO;2.
Newman, M., and Coauthors, 2016: The Pacific decadal oscillation, revisited. J. Climate, 29, 4399–4427, https://doi.org/10.1175/JCLI-D-15-0508.1.
Nigam, S., 2003: Teleconnections. Encyclopedia of Atmospheric Sciences, J. R. Holton, J. Pyle, and J. A. Curry, Eds., Academic Press, 2243–2269.
Nishii, K., H. Nakamura, and Y. J. Orsolini, 2011: Geographical dependence observed in blocking high influence on the stratospheric variability through enhancement and suppression of upward planetary-wave propagation. J. Climate, 24, 6408–6423, https://doi.org/10.1175/JCLI-D-10-05021.1.
Penny, S., G. H. Roe, and D. S. Battisti, 2010: The source of the midwinter suppression in storminess over the North Pacific. J. Climate, 23, 634–648, https://doi.org/10.1175/2009JCLI2904.1.
Reichler, T., J. Kim, E. Manzini, and J. Kröger, 2012: A stratospheric connection to Atlantic climate variability. Nat. Geosci., 5, 783–787, https://doi.org/10.1038/ngeo1586.
Reynolds, R. W., T. M. Smith, C. Liu, D. B. Chelton, K. S. Casey, and M. G. Schlax, 2007: Daily high-resolution-blended analyses for sea surface temperature. J. Climate, 20, 5473–5496, https://doi.org/10.1175/2007JCLI1824.1.
Richter, J. H., C. Deser, and L. Sun, 2015: Effects of stratospheric variability on El Niño teleconnections. Environ. Res. Lett., 10, 124021, https://doi.org/10.1088/1748-9326/10/12/124021.
Schemm, S., and T. Schneider, 2018: Eddy lifetime, number, and diffusivity and the suppression of eddy kinetic energy in midwinter. J. Climate, 31, 5649–5665, https://doi.org/10.1175/JCLI-D-17-0644.1.
Scott, R. K., and L. M. Polvani, 2004: Stratospheric control of upward wave flux near the tropopause. Geophys. Res. Lett., 31, L02115, https://doi.org/10.1029/2003GL017965.
Scott, R. K., and L. M. Polvani, 2006: Internal variability of the winter stratosphere. Part I: Time-independent forcing. J. Atmos. Sci., 63, 2758–2776, https://doi.org/10.1175/JAS3797.1.
Seager, R., Y. Kushnir, M. Visbeck, N. Naik, J. Miller, G. Krahmann, and H. Cullen, 2000: Causes of Atlantic Ocean climate variability between 1958 and 1998. J. Climate, 13, 2845–2862, https://doi.org/10.1175/1520-0442(2000)013<2845:COAOCV>2.0.CO;2.
Sigmond, M., J. F. Scinocca, V. V. Kharin, and T. G. Shepherd, 2013: Enhanced seasonal forecast skill following stratospheric sudden warmings. Nat. Geosci., 6, 98–102, https://doi.org/10.1038/ngeo1698.
Simpson, I. R., P. Hitchcock, T. G. Shepherd, and J. F. Scinocca, 2011: Stratospheric variability and tropospheric annular-mode timescales. Geophys. Res. Lett., 38, L20806, https://doi.org/10.1029/2011GL049304.
Simpson, I. R., P. Hitchcock, T. G. Shepherd, and J. F. Scinocca, 2013a: Southern annular mode dynamics in observations and models. Part I: The influence of climatological zonal wind biases in a comprehensive GCM. J. Climate, 26, 3953–3967, https://doi.org/10.1175/JCLI-D-12-00348.1.
Simpson, I. R., T. G. Shepherd, P. Hitchcock, and J. F. Scinocca, 2013b: Southern annular mode dynamics in observations and models. Part II: Eddy feedbacks. J. Climate, 26, 5220–5241, https://doi.org/10.1175/JCLI-D-12-00495.1.
Simpson, I. R., P. Hitchcock, R. Seager, Y. Wu, and P. Callaghan, 2018: The downward influence of uncertainty in the Northern Hemisphere stratospheric polar vortex response to climate change. J. Climate, 31, 6371–6391, https://doi.org/10.1175/JCLI-D-18-0041.1.
Sutton, R. T., W. A. Norton, and S. P. Jewson, 2001: The North Atlantic Oscillation—What role for the ocean? Atmos. Sci. Lett., 1, 89–100, https://www.sciencedirect.com/science/article/abs/pii/S1530261X00900185.
Taylor, K. E., D. Williamson, and F. Zwiers, 2000: The sea surface temperature and sea-ice concentration boundary conditions of AMIP II simulations. PCMDI Rep. 60, 20 pp., https://pcmdi.llnl.gov/report/ab60.html.
Wang, L., and W. Chen, 2014: An intensity index for the East Asian winter monsoon. J. Climate, 27, 2361–2374, https://doi.org/10.1175/JCLI-D-13-00086.1.
White, I. P., C. I. Garfinkel, E. P. Gerber, M. Jucker, P. Hitchcock, and J. Rao, 2020: The generic nature of the tropospheric response to sudden stratospheric warmings. J. Climate, 33, 5589–5610, https://doi.org/10.1175/JCLI-D-19-0697.1.
Yuval, J., H. Afargan, and Y. Kaspi, 2018: The relation between the seasonal changes in jet characteristics and the Pacific midwinter minimum in eddy activity. Geophys. Res. Lett., 45, 9995–10 002, https://doi.org/10.1029/2018GL078678.
Zhang, P., Y. Wu, I. R. Simpson, K. L. Smith, X. Zhang, B. De, and P. Callaghan, 2018: A stratospheric pathway linking a colder Siberia to Barents-Kara Sea sea ice loss. Sci. Adv., 4, eaat6025, https://doi.org/10.1126/sciadv.aat6025.
