Recent work has suggested that tropical Pacific decadal variability and external forcings have had a comparable influence on the observed changes in the Southern Hemisphere summertime eddy-driven jet over the satellite era. Here we contrast the zonally asymmetric response of the Southern Hemisphere eddy-driven jet to tropical Pacific decadal variability by designing an atmosphere-only PAC-A experiment using the Community Atmosphere Model version 5 (CAM5) and comparing it with the fully coupled Community Earth System Model Version 1 (CESM1) tropical Pacific pacemaker (PAC-C) experiments. In both frameworks, the tropical Pacific sea surface temperature (SST) anomalies are identical (model climatology plus observed anomalies), which allows the PAC-C and PAC-A experiments to be used to estimate the impact of coupling on teleconnections from the tropical Pacific to the Southern Hemisphere extratropics. The observed summertime South Pacific jet intensification is reproduced in both coupled and uncoupled experiments, indicating that the central and eastern tropical Pacific (hereafter, tropical Pacific) SST impacts the South Pacific jet mainly via direct atmospheric teleconnections. By contrast, only the coupled PAC-C captures the summertime poleward shift of the South Atlantic-Indian jet, suggesting that air-sea coupling is essential in driving the teleconnections between tropical Pacific SST anomalies and South Atlantic-Indian jet variations.

Key Points

Atmosphere-only experiments are designed to contrast with coupled pacemakers runs to understand the influence of tropical Pacific decadal variability on the Southern Hemisphere eddy-driven jet
The South Pacific jet responds to tropical Pacific sea surface temperature variability via direct atmospheric processes
Air-sea coupling is crucial for reproducing the observed poleward migration of the South Atlantic-Indian jet

Plain Language Summary

There is a zonally asymmetric response of the Southern Hemisphere summertime eddy-driven jet to tropical Pacific decadal variability, with a poleward migration observed in the South Atlantic-Indian basin and an intensification for the South Pacific basin. By designing an atmosphere-only experiment and comparing it with a fully coupled tropical Pacific pacemaker run, we find that the South Pacific jet responds to tropical Pacific sea surface temperature (SST) variability via direct atmospheric processes. By contrast, air-sea coupling is notably important for the South Atlantic-Indian jet response to tropical Pacific SSTs as it is conducive to the Pacific–South American Rossby wave propagation into the Atlantic.

1 Introduction

Observational and model studies have shown that the Southern Hemisphere summertime (December-January-February, DJF) eddy-driven jet experienced a poleward shift during the last four decades, along with a positive trend in the Southern Annular Mode (SAM). Along with the dominant impact of Southern Hemisphere stratospheric ozone depletion on these trends, as discussed in previous research (Banerjee et al., 2020; Lee & Feldstein, 2013; Polvani et al., 2011), the role of tropical sea surface temperature variations has also recently been emphasized (Deser & Phillips, 2009; Schneider et al., 2015; Yang et al., 2020). In particular, Yang et al. (2020) examined the separate influence of internally generated decadal variability from tropical Pacific, Indian and Atlantic basin sea surface temperatures (SSTs) on the position and strength of the Southern Hemisphere eddy-driven jet. Using the Community Earth System Model Version 1 (CESM1) pacemaker experiments (Deser et al., 2017), Yang et al. (2020) found the internally driven decadal changes of tropical Pacific SSTs (specifically the central and eastern tropical Pacific regions within the solid lines labeled in Figure 1c) and external forcing played comparable and dominant roles in the poleward jet shift between the 1979–1998 and 1999–2013 periods, with the tropical and North Atlantic and Indian Ocean SST making only a weak and sometimes offsetting contribution. Therefore, the focus of this study is on the dynamical mechanisms regulating the jet response to tropical Pacific SSTs.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

The decadal difference between averaged P2 (1999–2013) and P1 (1979–1998) for sea surface temperature (units: K) in the December-January-February (DJF) season. (a) Extended Reconstruction Sea Surface Temperature, version 3b (ERSSTv3b); (b) Community Earth System Model Version 1 (CESM) Large Ensemble mean (LENS), indicative of external forcing; (c) Pacific pacemaker ensemble-mean (PAC-C) minus LENS; (d) PAC-A ensemble-mean minus LENS; (e) extSPCZ-A ensemble-mean minus LENS. Stippling indicates differences are significant at the 95% level based on a t-test. In panels (c–e), within the solid black lines, the sea surface temperatures (SSTs) are nudged to the model climatology plus observed anomalies; between the solid and the dashed lines are the buffer belts where the nudging is linearly tapered to zero; outside of the dashed lines, the SSTs are free to evolve in PAC-C but prescribed with LENS SST in PAC-A and extSPCZ-A.

Yang et al. (2020) pointed out that in both the observations and in response to tropical Pacific SSTs, the summertime eddy-driven jet experiences an intensification in the South Pacific with a deepened Amundsen Sea Low, unlike the poleward movement in the South Atlantic and South Indian basins. These zonal asymmetric jet variations are also described by Waugh et al. (2020) for the DJFMAM seasons, that is, the South Pacific jet underwent a strengthening in contrast to the latitudinal position change in the South Atlantic-Indian basins. The tropical Pacific - South Pacific teleconnection has been widely discussed in previous studies, for example, Irving and Simmonds (2016) suggested that the interannual to interdecadal tropical SST trends over 1979–2014 can influence the Southern Hemisphere high latitudes and regional Antarctic climate via an anomalous Pacific–South American Rossby wave pattern. Meehl et al. (2016) linked the negative phase of the Interdecadal Pacific Oscillation (IPO) during 2000–2014 to the Ross Sea and Amundsen Sea sea ice changes via a positive phase of the SAM combined with an enhanced Amundsen Sea Low.

By contrast, the tropical Pacific remote impact on the South Atlantic/Indian atmospheric circulation has been less studied and previous work mainly attributes this remote connection to the downstream effect of the Pacific–South American pattern. For instance, Rodrigues et al. (2015) found in response to La Niña events and its induced positive Pacific–South American, a northeast-southwest South Atlantic SST dipole develops with a warming anomaly around 40°S and cooling anomaly at 20°S. Similar results are attained at the interdecadal timescales, where Lopez et al. (2016) found South Atlantic SST and sea surface height dipole variability is remotely modulated by the IPO via Pacific–South American wave trains.

The strength of using coupled Pacific pacemaker experiments in Yang et al. (2020) is that it allows coupled feedbacks in response to regionally prescribed SSTs, however, these pacemaker simulations cannot separate out the secondary atmospheric teleconnections that arise due to interbasin interactions, particularly between the tropical Pacific and adjacent oceans (Cai et al., 2019). Hence an outstanding question arises: Does tropical Pacific SST impact the jet variation via direct atmospheric pathway or by inducing SST changes in other basins first due to the air-sea coupling?

Yang et al. (2020) argue that the South Pacific jet intensification was primarily induced by internal fluctuations from tropical Pacific SST. However, other studies also suggest the South Pacific Convergence Zone (SPCZ) could potentially modulate the South Pacific mid-to-high latitude circulation response to tropical Pacific SST anomaly (Clem et al., 2019; Meehl et al., 2016, 2019). For example, using atmospheric heating experiments, Meehl et al. (2016) found that for the DJF season, the eastern Pacific cooling would induce a deepened Amundsen Sea Low while the SPCZ warming (170°E, 20°S) would lead to a shallowed Amundsen Sea Low instead, with the latter more consistent with observed changes in that season. Consistent results are obtained by Clem et al. (2019) for the DJFMAM season where a 2°C SST anomaly was prescribed over the SPCZ (160–175°E, 15–25°S) using the CAM5 model. However, both these studies apply idealized anomalies, and the separate influence of tropical eastern Pacific and SPCZ SST anomalies on the Southern Hemisphere circulation using realistic SSTs has yet to be explored.

There remain questions regarding the mechanism driving the zonal asymmetric South Pacific jet response to the tropical Pacific SST and the importance of coupled feedbacks to the jet response in general. Therefore, two main questions would be addressed in this study:

What is the role of air-sea coupling in the tropical Pacific SST and Southern Hemisphere eddy-driven jet teleconnections? What is the dynamical mechanism behind the coupling impact?
What are the relative influences of tropical eastern Pacific and SPCZ SST anomalies on the Amundsen Sea Low and eddy-driven jet variations?

The rest of this study is organized as follows. The model and experimental design is described in Section 2. The main results are presented in Section 3. A brief summary and discussion are outlined in Section 4.

2 Model and Experimental Design

In this study, the observed Zonal (U) and meridional (V) winds and mean sea level pressure datasets are taken from the European Centre for Medium-Range Weather Forecasts Interim reanalysis (ERA-Interim; Dee et al., 2011). The monthly SST data is from the National Oceanic and Atmospheric Administration (NOAA) Extended Reconstruction Sea Surface Temperature, version 3b (ERSSTv3b, Smith et al., 2008).

The fully coupled CESM1 40-member Large Ensemble (LENS) and 10-member Pacific pacemaker (PAC-C) experiments are identical to those employed in Yang et al. (2020), with LENS used to estimate the impact of external forcing, and PAC-C (with external forcing removed, that is, PAC-C minus LENS) reflecting the influence of internally driven tropical Pacific SST on the jet. Another two uncoupled simulations, PAC-A and extSPCZ-A, are designed to further separate the direct atmospheric teleconnection from tropical Pacific SST to the Southern Hemisphere jet (see Table 1).

Table 1. Summary of the CESM and CAM5 Experiments Used in This Study

Experiments	External forcing	SST forcing	Coupling	Members
Large Ensemble (LENS)	Historical for 1920–2005, RCP8.5 for 2006–2013	—	√	40
Pacific Pacemaker (PAC-C)	Same as LENS, except with SPARC ozone	Fully restored SSTA for 15°S–15°N, 175°W to the American coast.	√	10
PAC-A	Same as PAC-C	PAC-C SST for 15°S–15°N, 165°E to the American coast; LENS SST outside eastern Pacific.	×	10
extSPCZ-A	Same as PAC-C	PAC-C PAC-C SST for 35°S–15°N, 165°E to the American coast; LENS SST outside eastern Pacific.	×	10

The CESM Large Ensemble is generated by a small perturbation to the initial atmospheric temperature field at the beginning of the simulations in 1920, which then develops into a diverse spread and produces 40 different realizations of intrinsic climate variability. All LENS ensemble members are constrained to the same radiative forcing scenario (Taylor et al., 2012), with historical forcing during 1920–2005 and a high emission forcing scenario of the Representative Concentration Pathway (RCP) 8.5 (Moss et al., 2010) from 2006 to 2080 following the Coupled Model Intercomparison Project Phase 5 (CMIP5) design protocol. The LENS 40-member ensemble mean then averages out this random internal variability and reflects the impact of external radiative forcings on the climate system.

All the experiments listed in Table 1 use the same CESM1 code base and time-varying radiative forcing as the LENS, apart from a minor difference in ozone forcing. While the LENS uses the Whole Atmosphere Community Climate Model (WACCM, Marsh et al., 2013) ozone data set (see Kay et al., 2015), the PAC-C and uncoupled runs are forced with the Stratosphere-Troposphere Processes and their Role in Climate (SPARC) stratospheric ozone data (Cionni et al., 2011). However, according to Schneider and Deser (2017), the ozone forcing differences have statistically indistinguishable impacts on the trends in the Southern Hemisphere eddy-driven jet over the satellite era.

In the coupled pacemaker experiment (PAC-C), the SSTs in the tropical eastern Pacific are restored to the model climatology plus observed SST anomalies using a nudging technique (Kosaka & Xie, 2013). Specifically, the SST anomalies between 15°S and 15°N, 175°W to the American coast (shown as solid black lines in Figure 1c), are nudged to the observed SST anomalies from NOAA ERSSTv3b. There are three buffer belts along 15°S–20°S, 15°N–20°N and 175°W–180° (dashed black lines in Figure 1c), where the SST anomalies are linearly tapered to zero. In this way, as described in Yang et al. (2020), the PAC-C contains both the radiatively forced and internally driven variability from observed tropical Pacific SSTs. Subtracting LENS from PAC-C then removes the externally forced signal and provides an estimate of the response of the global climate system to the observed time-varying internally generated tropical Pacific SSTs.

However, in the coupled PAC-C where the SST anomalies over the eastern Pacific region are nudged to observations and the rest of the model is free to evolve (Table 1 and Figure 1c), the tropical Pacific SST forcing promotes an SST response in other regions, which can then force a secondary teleconnection and influence the mid-latitude jet. This means in the PAC-C set, the tropical Pacific SST can influence the jet in two ways, one is via a direct atmospheric impact, the other is via the Pacific inducing SST changes in adjacent ocean basins first, and then having a secondary impact on the jet.

To remove the compounding effects of Pacific coupling to other ocean basins, an atmosphere-only simulation (hereafter denoted PAC-A) is designed to compare with the PAC-C. This uncoupled PAC-A experiment is intended to match the coupled PAC-C in every way, apart from the inclusion of coupling outside of the tropical Pacific region. The PAC-A runs use the same atmospheric model (CAM5) as in the PAC-C and are subjected to identical external forcing. The PAC-A SST forcing in the tropical eastern Pacific region is prescribed using the time-evolving PAC-C SSTs (i.e., model climatology plus observed anomalies) from 165°E to the American coast. The western boundary of this prescribed region (solid box in Figure 1d) is extended by 20-degree west to 165°E, compared to the nudged region in the PAC-C (solid box in Figure 1c) experiment which stops at 175°W. This extended region better mimics the coupled PAC-C experiment which naturally extends the anomalies related to tropical eastern Pacific oscillations (such as the IPO and El Niño–Southern Oscillation (ENSO)) to the west of the nudged region due to coupled air-sea dynamics.

Outside of the tropical Pacific, the time-evolving LENS 40-member ensemble mean SST is imposed. In this way, the PAC-A 10-member ensemble mean with external forcing removed (i.e., PAC-A minus LENS) isolates the direct atmospheric influence of tropical Pacific SST on the Southern Hemisphere atmospheric circulation which is of interest here.

Another uncoupled experiment (hereafter named extSPCZ-A), with PAC-C SSTs applied over 35°S–20°N, 165°E to the American coast (Figure 1e), is also performed to further investigate the relative role of observed tropical eastern Pacific and SPCZ SST anomalies. Since PAC-A and extSPCZ-A are identical except for the expanded nudged region, comparisons between extSPCZ-A and PAC-A distinguishes the impact of the SPCZ SST variations on the atmospheric teleconnection patterns in the SH mid-to-high latitudes.

3 Results

3.1 Tropical Pacific Pacemaker and PAC-A Comparison

Based on the IPO phase transition (positive to negative) around 1999, and to follow Yang et al. (2020), we separate the satellite era into two periods 1980–1998 (P1) and 1999–2013 (P2). Figure 1 shows the decadal difference (P2 average minus P1 average) of SST for observations and model simulations. In the tropical eastern and central Pacific, the observations, PAC-C and PAC-A SST all show large cooling anomalies, which is typical of the negative phase of the IPO (Figures 1a, 1c and 1d). Outside of this region, the coupled pacemaker response is a result of (a) the atmospheric response to the nudged tropical Pacific SST and (b) the secondary impact from SST anomalies in other basins, that have evolved in response to the tropical Pacific SSTs. Specifically, in PAC-C minus LENS (Figure 1c), the eastern Pacific cooling promotes a warming over the SPCZ and the South Atlantic Convergence Zone. This is consistent with Clem et al. (2019) who argued that cooling in the eastern Pacific can induce a warming anomaly in the SPCZ region as part of the negative IPO pattern. The negative IPO induced South Atlantic warming anomaly in Figure 1c is also consistent with Lopez et al. (2016) in terms of the decadal teleconnections between the negative IPO and South Atlantic SST changes. By contrast, PAC-A isolates the direct atmospheric impact of tropical Pacific SST anomalies alone.

In terms of summertime variations, a poleward shift of the eddy-driven jet is observed in the Atlantic and Indian basins and an in-place intensification over the Pacific basin (Figure 2a). As discussed in Yang et al. (2020), external forcing (Figure 2b) drives an annular poleward displacement of the eddy-driven jet, with the largest impact in the South Atlantic basin. Internal variability in tropical Pacific SST, on the other hand, contributes to both the poleward shift of the South Atlantic-Indian jet as well as the strengthening of the South Pacific jet (Figure 2c, PAC-C with external forcing removed). The distinct zonally asymmetric jet strengthening at 45°S in the South Pacific basin is reproduced in the PAC-A experiment (Figure 2d, with external forcing removed). However, the PAC-A experiment fails to capture the poleward displacement of the jet over the South Atlantic-Indian basins, exhibiting non-significant weakening over the Atlantic-Indian sector. Differences between the uncoupled PAC-A and coupled PAC-C implies that the influence of tropical Pacific SST on South Pacific jet intensification is mainly via atmospheric processes. It is also clear that for the teleconnections between the tropical Pacific SST and South Atlantic-Indian jet variations, air-sea coupling is critical since the observed poleward shift of the jet in those basins is only reproduced in the coupled PAC-C experiment.

These zonally asymmetric results are also found in the sea level pressure field. In the South Atlantic-Indian sector a positive SAM phase is observed, with high-pressure anomaly at 40°S and low-pressure anomaly along 65°S (Figure 3a). This pattern is simulated in the PAC-C experiment (Figure 3c), consistent with a poleward displacement of the jet. In contrast, PAC-A exhibits a non-significant negative SAM-like pattern with weakened jet, largely diverging from the observations, which again indicates that coupling is important to establish the teleconnections between the tropical Pacific and the South Atlantic-Indian basin. In the South Pacific basin, the deepened Amundsen Sea Low and 30°S high-pressure anomaly is well captured in both the PAC-C and PAC-A experiments.

This tropical-extratropical teleconnection within the Pacific basin is also detected in the other three seasons. In the reanalysis, the largest decadal trend of the South Pacific jet has occurred during the MAM season (Schneider et al., 2015), with significant enhancement over 40–50°S, 170°E−70°W. In accordance with this strengthened South Pacific jet during austral autumn, a deepened Amundsen Sea Low has also been observed (not shown). This trend is largely reproduced by the PAC-A (with external forcing removed, Figure 4c), highlighting the role of the direct atmospheric pathway in this tropical Pacific-South Pacific teleconnection (the SST differences in MAM are similar to DJF, not shown). This builds on Schneider et al. (2015), who attributed the South Pacific changes in MAM to tropical SSTs using experiments with CAM4, by narrowing the attribution to tropical Pacific SSTs specifically. By contrast, the South Pacific jet strengthening in austral autumn is not well captured by the coupled PAC-C ensemble mean (with external forcing removed, Figure 4b), indicating that air-sea coupling might induce other secondary circulation changes that offset the Rossby wave propagation from the central equatorial Pacific to the Amundsen Sea Low.

In the austral winter-spring seasons (June-July-August, JJA, and September-October-November, SON), the observed South Pacific eddy-driven jet exhibits an equatorward shift, with the peak westerly increase centered at 40–45°S, where both the PAC-A and PAC-C ensemble mean captures this pattern. The South Pacific jet strengthening during DJF-MAM, as well as the equatorward displacement during JJA-SON, are in agreement with Waugh et al. (2020), who suggest that the westerly trends in JJA-SON are largely due to the internal variability since the reanalysis fall within the ensemble spread for historical climate model simulations. Based on our PAC-A results, we further find that the South Pacific westerly variation in all seasons is mainly induced by direct atmospheric teleconnections from tropical Pacific SST anomalies. In the non-summer seasons, the South Atlantic-Indian jet exhibits less uniform changes in both intensity and position compared to the DJF season, and neither PAC-C nor PAC-A can fully reproduce the South Atlantic-Indian jet variation, indicating there are other factors beyond tropical Pacific SST anomaly that drive the South Atlantic-Indian jet changes outside of the summertime. Our results are also consistent with the observational study of Clem et al. (2017), who found that eastern tropical Pacific SST was significantly related to the South Pacific jet in all seasons (in particular during DJF-MAM the South Pacific jet strengthening is tied to direct atmospheric forcing associated with the eastern tropical Pacific cooling), while in the Atlantic and Indian sectors, the eastern tropical Pacific cooling had much weaker relationships with the jet variations outside of the DJF season.

3.2 PAC-A and extSPCZ-A Comparison

Differences between the PAC-A and PAC-C in the South Atlantic-Indian jet simulations indicate that tropical Pacific SSTs could influence the Southern Hemisphere jet by modifying SSTs outside this region which then induce atmospheric teleconnections. Based on Figure 1c, the eastern Pacific SST cooling develops a warming belt in the Southern Hemisphere subtropical area in each basin, including the SPCZ, South Atlantic and Indian convergence zones. Previous studies (Clem et al., 2019; Meehl et al., 2016) suggest that SPCZ warming shallows the Amundsen Sea Low in the DJF-MAM seasons, playing an opposite role to eastern Pacific SST cooling.

Accordingly, another experiment is conducted to further investigate the relative roles of eastern Pacific SST cooling and its induced warming in the southeast SPCZ as seen in Figures 1c and 1e. Comparison between extSPCZ-A and PAC-A (Figures 2 and 7e) shows little difference in the South Pacific jet variations during austral summer, which suggests that the eastern Pacific cooling is the main driver of the South Pacific jet strengthening, and that contributions from the SPCZ warming are rather small. Similar relations can be found in the other three seasons (Figures 4-6). The extSPCZ-A captures a slight poleward displacement of the South Indian jet during DJF, which is in better agreement with the PAC-C results than PAC-A, suggesting that SPCZ warming may be important to enable the tropical Pacific SST influence on the South Indian jet variations. However, in the South Atlantic basin, neither PAC-A nor extSPCZ-A reproduce the southward movement in the jet that can be seen in observations and in the PAC-C.

4 Mechanisms

4.1 Poleward Propagation of Rossby Waves

Early studies imply that tropical Pacific SSTs can impact the extratropical circulation via the poleward propagation of stationary Rossby waves (Hoskins & Karoly, 1981), and recent work further suggests the SPCZ SST anomaly as an additional hotspot for Rossby wave sources (Clem et al., 2019; Lopez et al., 2016). Here we examine this teleconnection at the decadal timescales in the uncoupled runs and compare with the coupled PAC-C to identify the role of air-sea coupling.

For the observed SST decadal difference, there is a cooling of the central tropical Pacific SST as noted earlier (Figure 1a), which leads to reduced precipitation with descending air (Figure 8a, the largest negative values occur on and just south of the Equator, at around 150°E), and more outgoing longwave radiation (OLR) over that same area (Figure 8b). The positive OLR anomaly coincides with an upper troposphere horizontal convergence in the central Pacific (negative geopotential height anomaly at 200 hPa, Figure 8c). Mirroring the tropical Pacific cooling, there is also significant warming over the SPCZ near 30°S. This region shows opposite tendencies compared to the tropical central Pacific, with enhanced rainfall, less OLR and a positive 200 hPa geopotential height anomaly. These two centers (central Pacific cooling and SPCZ warming), can both act as Rossby wave sources for the extratropics (Clem et al., 2019), driving a positive geopotential height anomaly at around 40° S and a negative tendency along 65° S at 200 hPa (Figure 8c). Since the mid-latitude atmospheric circulation is equivalent barotropic, at the surface level the observed Southern Hemisphere circulation response (with external forcing included) also shows positive sea level pressure anomalies over mid-latitudes and negative sea level pressure anomalies along high-latitudes, resulting in a positive SAM phase (Figure 3a).

The PAC-C (with external forcing removed) captures most features in the observations, with a cooling SST anomaly, associated decreased rainfall, positive OLR and upper-level convergence in the central Pacific, as well as a warming SST anomaly, enhanced rainfall, negative OLR and upper-level divergence over the SPCZ (Figures 8d–8f). However, the Rossby wave train produced by the PAC-C (external forcing removed) is more prominent and zonally asymmetric compared to observations, with distinct low-high geopotential height fluctuations propagating from the tropical Pacific to the subtropical South Atlantic via South America, forming a clear Pacific–South American Rossby wave pattern (Figure 8f). This suggests a strong role for internally driven tropical Pacific SST variability in driving the asymmetric component of the observed response.

To better depict the evolution of the Pacific–South American pattern, here we examine the Rossby wave activity flux, which can be used to diagnose the propagation of stationary wave energy on a zonal background flow. Following (Wang et al., 2019), we calculate the horizontal component of the Rossby wave flux on the 200 hPa, as:

$urn:x-wiley:2169897X:media:jgrd57363:jgrd57363-math-0001$ (1)

where $urn:x-wiley:2169897X:media:jgrd57363:jgrd57363-math-0002$ is the background wind field; and $urn:x-wiley:2169897X:media:jgrd57363:jgrd57363-math-0003$ is the streamfunction anomaly. The zonal background flow is the DJF wind climatology over 1980–2013.

In the reanalysis, the wave activity flux emanates from the central subtropical Pacific (15–20°S, 150°W) as well as directly to the east of Australia and propagates poleward to the South Atlantic basin. In PAC-C, in response to the tropical Pacific cooling at the Equator (i.e., negative IPO), there is an outflow of the stationary wave activity flux originating from 15 to 20°S, 120°W, co-located with a low-pressure anomaly (cyclonic tendency) at upper levels. This wave activity flux propagates poleward and eastward toward the Antarctic Peninsula and circulates back to the tropics in the South Atlantic basin following a great circle track (Figure 8f). The wave activity flux in PAC-C has a stronger Rossby wave train compared to observations, with the latter including the externally forced response – not just the response to the negative IPO.

The PAC-A (external forcing removed) reproduces the observed rainfall and OLR anomalies in the tropical Pacific. The 200 hPa geopotential height field displays a mid-latitude anticyclonic anomaly and deepened Amundsen Sea Low in the South Pacific basin, aligned with the wave activity flux, which travels southeastward toward the southern tip of South America, and then dissipates after turning around the south of South America (Figure 9c). In other words, the Pacific–South American pattern and the wave activity flux in PAC-A is well resolved within the South Pacific basin but vanishes upon entering the South Atlantic region. Comparisons between PAC-A and PAC-C thus highlight the importance of coupling in the Rossby wave propagation and in maintaining the tropical Pacific-South Atlantic-Indian teleconnections.

The combined impact of tropical Pacific cooling and SPCZ warming (Figure 9f) exhibits a similar Rossby wave pattern to PAC-A in the South Pacific basin, although the deepened low shifts slightly east from the Ross Sea (in PAC-A) to the Amundsen Sea (in extSPCZ-A), which is in closer agreement with the PAC-C simulations. The wave activity flux in the mid-latitude South Pacific shows similar results, with comparable southeastward flow as in the PAC-C, but the magnitude of the wave activity flux in the South Atlantic is much reduced. In the South Indian basin, the positive SAM-like pattern is better captured in extSPCZ-A, indicating SPCZ warming could be bridging the teleconnections between tropical SST anomaly and South Indian jet variations. The summertime Amundsen Sea Low longitudinal location is better simulated with the additive influence of tropical Pacific cooling and southeast SPCZ warming, however its intensity does not vary much with/without SPCZ warming (Figure 3e). This is different to the findings of Clem et al. (2019), who suggest that during the DJFMAM season, the west SPCZ warming plays an opposite role to the eastern Pacific cooling by shallowing the Amundsen Sea Low. This is likely due to the different anomalies applied in the SPCZ region between our experiments and those of Clem et al. (2019). While our observation-derived extSPCZ-A experiments have significant anomalies primarily in the south-eastern portion of the SPCZ, Clem et al. (2019) focused on the role of the west SPCZ warming via a +2°C SST anomaly. If the combined impact of SPCZ warming and tropical Pacific cooling is not linearly additive, this idealized SST forcing (which is considerably larger than the observed trend) could cause a significant difference from the observations. Also, Clem et al. (2019) note that SSTs in the west part of the SPCZ are the key to generating a Rossby wave that propagates into the south-eastern Pacific in the DJF season, due to being closer to the exit region of the weak subtropical jet. The lack of observed internally driven anomalies in this location could explain why our extSPCZ-A experiment does not capture the Rossby wave train originating from the sub-tropical jet exit region, distinct from what was found in Clem et al. (2019).

4.2 Eddy-Mean Flow Interactions

Earlier studies indicate the tropical Pacific SST can influence the extra-tropical circulation via modulating the eddy momentum fluxes along the eddy-driven jet (Clem et al., 2017; Fogt et al., 2011; Yang et al., 2020). For example, in Yang et al. (2020) the influence of tropical Pacific SST on the zonal-mean westerlies was explained via the meridional temperature contrast and associated transient eddy-mean flow interactions. In that study, it was found that in both observations and the coupled pacemaker, a significant air temperature warming anomaly could be found throughout the troposphere at 45°S in response to anomalously cool tropical Pacific SSTs, which then lead to an enhanced meridional temperature gradient, increased near-surface baroclinity, more transient eddy generation (eddy momentum flux convergence anomaly) at high latitudes and a poleward shifted eddy-driven jet.

We also examine the air temperature distribution in the uncoupled experiments (Figure 10), but instead of zonal mean metrics as analyzed in Yang et al. (2020), here we examine the South Pacific and Atlantic-Indian basins separately, given our focus on the jet’s zonally asymmetric variations. In the South Pacific basin, both PAC-A and extSPCZ-A reproduce a significant warming anomaly throughout the troposphere as in the PAC-C, although with a diminished magnitude (Figures 10c and 10d). As a consequence of the enhanced meridional temperature contrast at around 50°S, more transient eddies would be generated at high latitudes, which could be potentially related to the substantial jet intensification in the South Pacific basin in both observations and all model simulations.

Over the South Atlantic-Indian region, in contrast to the extensive warming belt in the PAC-C, PAC-A simulates only a very weak warming anomaly in the mid-latitudes (Figure 10g), leading to a depressed tropic-to-pole temperature contrast, less baroclinity and eddy generation, and the absence of the poleward displacement of the Atlantic-Indian jet at 50°S. The South Atlantic-Indian temperature contrast and jet displacement in extSPCZ-A lie between PAC-A and the PAC-C, that is, the warming band in extSPCZ-A is larger than in PAC-A, but much weaker compared to PAC-C, resulting in only a moderate meridional temperature gradient and partial poleward movement of the Indian Ocean jet.

The absence of air temperature gradients across the Atlantic-Indian mid-latitudes in PAC-A could be related to the SST patterns discussed above. In the coupled PAC-C, responding to the tropical Pacific cooling, the Southern Hemisphere subtropical SST and the tropospheric temperature aloft involves a warming anomaly in each basin (i.e., South Pacific, Atlantic and Indian convergence zones at the surface) and prompts an increased meridional temperature gradient at about 50°S. In contrast, the tropospheric temperature warming anomaly at 40°S is considerably less significant in the PAC-A in the absence of subtropical SST warming anomalies, which further leads to disparities in the latitudinal position of the Atlantic-Indian jet between the PAC-A and PAC-C simulations.

5 Summary and Discussions

Recent work has suggested that the IPO has made a major contribution to the Southern Hemisphere summertime eddy-driven jet variations over the satellite era, including the strengthening of the South Pacific jet and poleward shift of the South Atlantic-Indian jet (Yang et al., 2020). One of the remaining questions, however, is whether tropical Pacific SSTs influence the Southern Hemisphere jet via direct atmospheric teleconnections or by interbasin interactions which first change the SSTs in other basins. In this study, we addressed this question by designing an atmosphere-only PAC-A experiment using CAM5 and comparing it with the fully coupled CESM1 tropical Pacific pacemaker experiments (referred to here as PAC-C; as analyzed in Yang et al., 2020). In both frameworks, the tropical Pacific SST anomalies follow the observed trajectory, which allows the PAC-C and PAC-A member ensemble mean (with external forcing excluded, i.e., minus LENS) to estimate the response to the internally driven tropical Pacific SST variability. Outside of the tropical Pacific, the SST in the PAC-C can freely respond to the tropical Pacific through interbasin interactions, whereas in the uncoupled PAC-A runs, SSTs are constrained to follow the LENS ensemble mean SST. Therefore, the PAC-A minus LENS can isolate the direct atmospheric impact of tropical Pacific SST on Southern Hemisphere eddy-driven jet changes.

We found that in all seasons, the uncoupled PAC-A could capture the observed eddy-driven jet variation in the South Pacific basin, including the equatorward shift in the JJA-SON seasons and strengthening during DJF-MAM. This indicates that the South Pacific eddy-driven jet responds to tropical Pacific SST mainly via atmospheric pathways. Comparisons between PAC-C and PAC-A suggest that the air-sea coupling is notably more important in forcing the South Atlantic-Indian basin to enable the Pacific–South American Rossby wave propagation. A further uncoupled experiment extSPCZ-A was conducted to investigate the relative roles of tropical Pacific and SPCZ convective centers for the above teleconnections. Differences between PAC-A and extSPCZ-A reveal the importance of the SPCZ SST warming in bridging the Pacific–South American Rossby wave pattern and generating a more realistic Amundsen Sea Low position and South Indian jet variations. The basin-averaged meridional temperature gradient in each model experiment is also analyzed to explain the jet similarities in the South Pacific basin and disparities in the South Atlantic-Indian basin between PAC-C and PAC-A simulations.

Waugh et al. (2020) suggested that the austral winter-spring trends in the latitude of the Pacific jet are due to internal atmospheric variability based on the fact that the observed trends lie within the CMIP5 ensemble member spread. In this study, using the CESM1 LENS, coupled PAC-C and uncoupled PAC-A simulations, we systematically analyzed the impact of internally driven tropical Pacific SST variability on the westerlies in all seasons, and further clarified which internal process (i.e., direct atmospheric pathway or coupling) is predominantly important for different basins. We concluded that tropical Pacific SST influences the South Pacific jet decadal variation mainly via a direct atmospheric pathway in all seasons, but for the South Atlantic-Indian basins, the air-sea coupling is more important, especially in the DJF season.

Despite the Pacific–South American Rossby wave propagation, another possible dynamic pathway for tropical Pacific SST influence on the South Atlantic-Indian jet could be the tropical Pacific and Atlantic inter-basin interactions. For example, in response to the tropical Pacific cooling, the tropical Atlantic evolves a warming SST anomaly in observation (Figure 1a), which could further impact the Southern Hemisphere mid-latitude jet by the Hadley Cell circulation changes or via Rossby waves (Li et al., 2015; Simpkins et al., 2014). These potential mechanisms worth to be further addressed in the future.

Transient eddy-mean flow interactions have also been compared in the coupled PAC-C and uncoupled PAC-A simulations to explain the difference in the latitudinal position of the Atlantic-Indian jet between these two simulations. In observations and in PAC-C, the South Atlantic-Indian jet moves poleward in association with increased mid-latitude temperatures at around 45°S (Figures 10e and 10f) which enhances the meridional temperature gradient. By contrast, in PAC-A, in the absence of the mid-latitude SST warming anomalies over the South Atlantic and Indian Ocean (Figure 1d), the mid-latitude temperature warming anomaly in the troposphere is missing (Figure 10g), leading to a degraded tropic-to-pole temperature contrast. As a result, the PAC-A fails to reproduce the poleward displacement of the Atlantic-Indian jet at 50°S.

Previous studies have found a poleward jet shift can result from imposed temperature anomalies in the subtropical and mid-latitudes (Tandon et al., 2013), however jet shifts themselves can also induce adiabatic warming in the middle troposphere (Lim et al., 2016; Yang et al., 2020). In other words, the warming anomalies shown in observations and in PAC-C (Figures 10e and 10f) include the impact of SST warming from the subtropical South Atlantic-Indian Ocean and any secondary adiabatic heating generated by the poleward-shifted eddy-driven jet. Nevertheless, this positive feedback between the jet shift and middle troposphere warming has little impact on the key conclusion that coupling is required to reproduce the observed poleward shift in the South Atlantic-Indian jet in response to tropical Pacific SSTs.

Our findings could provide useful references for future air-sea-ice coupling studies. For instance, previous studies (Holland et al., 2017; Lefebvre et al., 2004) have found that variations in the mid-latitude jet can have a large impact on the Southern Ocean circulation and Antarctic sea ice distribution. Therefore, applying a more realistic zonally asymmetric wind forcing, as suggested in this study, instead of zonal-mean SAM-like winds for ocean sea-ice models, could potentially lead to more accurate sea ice predictions (Goyal et al., 2021), especially over the Amundsen and Ross Sea regions. In addition, the uncoupled PAC-A and coupled PAC-C simulations enable an “apple-to-apple” comparison to identify the tropical Pacific SST influence on global SST and atmospheric teleconnections. Understanding the role of tropical internal variability on the Southern Hemisphere eddy-driven jet is critical to understanding past and future changes in the Southern Ocean circulation and Antarctic sea-ice and ice sheets.

Acknowledgments

This work was supported by (a) the Australian Research Council Centre of Excellence for Climate Extremes (grant CE170100023); (b) the Regional and Global Model Analysis (RGMA) component of the Earth and Environmental System Modeling Program of the U.S. Department of Energy's Office of Biological & Environmental Research (BER) Cooperative Agreement # DE-FC02-97ER62402; (c) the National Center for Atmospheric Research, which is a major facility sponsored by the National Science Foundation under Cooperative Agreement No. 1852977. The authors thank the CESM Climate Variability and Change Working Group for providing the CESM1 Large Ensemble and pacemaker simulations which are available on the Earth System Grid (www.earthsystemgrid.org/). The authors thank Adam Phillips and Nan Rosenbloom (from NCAR), Shayne McGregor and Giovanni Liguori (from Monash University), and Harry Hendon (from the Australian Bureau of Meteorology) for helpful discussions. The authors also thank Kyle Clem and two anonymous reviewers for their helpful comments, which improved this manuscript.

Open Research

Data Availability Statement

Data sets of the uncoupled experiments can be downloaded from https://doi.org/10.5065/xdqz-6920. ERA-Interim reanalysis data are from https://apps.ecmwf.int/datasets/data/interim-full-moda. NOAA ERSSTv3b data are available from https://www.ncdc.noaa.gov/data-access/marineocean-data/extended-reconstructed-sea-surface-temperature-ersst-v3b. NOAA Interpolated Outgoing Longwave Radiation (OLR) can be found at https://psl.noaa.gov/data/gridded/data.interp_OLR.html. Global Precipitation Climatology Project (GPCP) precipitation data are available from http://www.esrl.noaa.gov/psd/data/gridded/data.gpcp.html. Most figures were created using NCAR Command Language (NCL) version 6.5.0. Computing resources (doi:10.5065/D6RX99HX) were provided by the Climate Simulation Laboratory at NCAR's Computational and Information Systems Laboratory.

References

Banerjee, A., Fyfe, J. C., Polvani, L. M., Waugh, D., & Chang, K.-L. (2020). A pause in southern hemisphere circulation trends due to the montreal protocol. Nature, 579(7800), 544–548. https://doi.org/10.1038/s41586-020-2120-4
Cai, W., Wu, L., Lengaigne, M., Li, T., McGregor, S., Kug, J.-S., et al. (2019). Pantropical climate interactions. Science, 363(6430), eaav4236. https://doi.org/10.1126/science.aav4236
Cionni, I., Eyring, V., Lamarque, J.-F., Randel, W., Stevenson, D., Wu, F., et al. (2011). Ozone database in support of CMIP5 simulations: Results and corresponding radiative forcing. Atmospheric Chemistry and Physics, 11(21), 11267–11292. https://doi.org/10.5194/acp-11-11267-2011
Clem, K. R., Lintner, B. R., Broccoli, A. J., & Miller, J. R. (2019). Role of the south Pacific convergence zone in west Antarctic decadal climate variability. Geophysical Research Letters, 46, 6900–6909. https://doi.org/10.1029/2019gl082108
Clem, K. R., Renwick, J. A., & McGregor, J. (2017). Relationship between eastern tropical Pacific cooling and recent trends in the southern hemisphere zonal-mean circulation. Climate Dynamics, 49(1–2), 113–129. https://doi.org/10.1007/s00382-016-3329-7
Dee, D. P., Uppala, S. M., Simmons, A., Berrisford, P., Poli, P., Kobayashi, S., et al. (2011). The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137(656), 553–597. https://doi.org/10.1002/qj.828
Deser, C., & Phillips, A. S. (2009). Atmospheric circulation trends, 1950–2000: The relative roles of sea surface temperature forcing and direct atmospheric radiative forcing. Journal of Climate, 22(2), 396–413. https://doi.org/10.1175/2008jcli2453.1
Deser, C., Simpson, I. R., McKinnon, K. A., & Phillips, A. S. (2017). The northern hemisphere extratropical atmospheric circulation response to ENSO: How well do we know it and how do we evaluate models accordingly? Journal of Climate, 30(13), 5059–5082. https://doi.org/10.1175/jcli-d-16-0844.1
Fogt, R. L., Bromwich, D. H., & Hines, K. M. (2011). Understanding the SAM influence on the south Pacific ENSO teleconnection. Climate Dynamics, 36(7–8), 1555–1576. https://doi.org/10.1007/s00382-010-0905-0
Goyal, R., Sen Gupta, A., Jucker, M., & England, M. H. (2021). Historical and projected changes in the southern hemisphere surface westerlies. Geophysical Research Letters, 48(4), e2020GL090849. https://doi.org/10.1029/2020gl090849
Holland, M. M., Landrum, L., Kostov, Y., & Marshall, J. (2017). Sensitivity of Antarctic sea ice to the southern annular mode in coupled climate models. Climate Dynamics, 49(5–6), 1813–1831. https://doi.org/10.1007/s00382-016-3424-9
Hoskins, B. J., & Karoly, D. J. (1981). The steady linear response of a spherical atmosphere to thermal and orographic forcing. Journal of the Atmospheric Sciences, 38(6), 1179–1196. https://doi.org/10.1175/1520-0469(1981)038<1179:tslroa>2.0.co;2
Irving, D., & Simmonds, I. (2016). A new method for identifying the Pacific–South American pattern and its influence on regional climate variability. Journal of Climate, 29(17), 6109–6125. https://doi.org/10.1175/jcli-d-15-0843.1
Kay, J. E., Deser, C., Phillips, A., Mai, A., Hannay, C., Strand, G., et al. (2015). The community earth system model (CESM) large ensemble project: A community resource for studying climate change in the presence of internal climate variability. Bulletin of the American Meteorological Society, 96(8), 1333–1349. https://doi.org/10.1175/bams-d-13-00255.1
Kosaka, Y., & Xie, S.-P. (2013). Recent global-warming hiatus tied to equatorial Pacific surface cooling. Nature, 501(7467), 403–407. https://doi.org/10.1038/nature12534
Lee, S., & Feldstein, S. B. (2013). Detecting ozone-and greenhouse gas–driven wind trends with observational data. Science, 339(6119), 563–567. https://doi.org/10.1126/science.1225154
Lefebvre, W., Goosse, H., Timmermann, R., & Fichefet, T. (2004). Influence of the southern annular mode on the sea ice–ocean system. Journal of Geophysical Research, 109(C9). https://doi.org/10.1029/2004jc002403
Li, X., Gerber, E. P., Holland, D. M., & Yoo, C. (2015). A Rossby wave bridge from the tropical Atlantic to West Antarctica. Journal of Climate, 28(6), 2256–2273. https://doi.org/10.1175/jcli-d-14-00450.1
Lim, E.-P., Hendon, H. H., Arblaster, J. M., Chung, C., Moise, A. F., Hope, P., et al. (2016). Interaction of the recent 50 year SST trend and La Niña 2010: Amplification of the southern annular mode and Australian springtime rainfall. Climate Dynamics, 47(7–8), 2273–2291. https://doi.org/10.1007/s00382-015-2963-9
Lopez, H., Dong, S., Lee, S. K., & Campos, E. (2016). Remote influence of interdecadal Pacific Oscillation on the South Atlantic meridional overturning circulation variability. Geophysical Research Letters, 43(15), 8250–8258. https://doi.org/10.1002/2016gl069067
Marsh, D. R., Mills, M. J., Kinnison, D. E., Lamarque, J.-F., Calvo, N., & Polvani, L. M. (2013). Climate change from 1850 to 2005 simulated in CESM1 (WACCM). Journal of Climate, 26(19), 7372–7391. https://doi.org/10.1175/jcli-d-12-00558.1
Meehl, G. A., Arblaster, J. M., Bitz, C. M., Chung, C. T. Y., & Teng, H. (2016). Antarctic sea-ice expansion between 2000 and 2014 driven by tropical Pacific decadal climate variability. Nature Geoscience, 9, 590–595. https://doi.org/10.1038/ngeo2751
Meehl, G. A., Arblaster, J. M., Chung, C. T., Holland, M. M., DuVivier, A., Thompson, L., et al. (2019). Sustained ocean changes contributed to sudden Antarctic sea ice retreat in late 2016. Nature Communications, 10(1), 14. https://doi.org/10.1038/s41467-018-07865-9
Moss, R. H., Edmonds, J. A., Hibbard, K. A., Manning, M. R., Rose, S. K., Van Vuuren, D. P., et al. (2010). The next generation of scenarios for climate change research and assessment. Nature, 463(7282), 747–756. https://doi.org/10.1038/nature08823
Polvani, L. M., Waugh, D. W., Correa, G. J., & Son, S.-W. (2011). Stratospheric ozone depletion: The main driver of twentieth-century atmospheric circulation changes in the southern hemisphere. Journal of Climate, 24(3), 795–812. https://doi.org/10.1175/2010jcli3772.1
Rodrigues, R. R., Campos, E. J., & Haarsma, R. (2015). The impact of ENSO on the south Atlantic subtropical dipole mode. Journal of Climate, 28(7), 2691–2705. https://doi.org/10.1175/jcli-d-14-00483.1
Schneider, D. P., & Deser, C. (2017). Tropically driven and externally forced patterns of Antarctic sea ice change: Reconciling observed and modeled trends. Climate Dynamics.
Schneider, D. P., Deser, C., & Fan, T. (2015). Comparing the impacts of tropical SST variability and polar stratospheric ozone loss on the southern ocean westerly winds. Journal of Climate, 28(23), 9350–9372. https://doi.org/10.1175/jcli-d-15-0090.1
Simpkins, G. R., McGregor, S., Taschetto, A. S., Ciasto, L. M., & England, M. H. (2014). Tropical connections to climatic change in the extratropical Southern Hemisphere: The role of Atlantic SST trends. Journal of Climate, 27(13), 4923–4936. https://doi.org/10.1175/jcli-d-13-00615.1
Smith, T. M., Reynolds, R. W., Peterson, T. C., & Lawrimore, J. (2008). Improvements to NOAA’s historical merged land–ocean surface temperature analysis (1880–2006). Journal of Climate, 21(10), 2283–2296. https://doi.org/10.1175/2007jcli2100.1
Tandon, N. F., Gerber, E. P., Sobel, A. H., & Polvani, L. M. (2013). Understanding Hadley cell expansion versus contraction: Insights from simplified models and implications for recent observations. Journal of Climate, 26(12), 4304–4321. https://doi.org/10.1175/jcli-d-12-00598.1
Taylor, K. E., Stouffer, R. J., & Meehl, G. A. (2012). An overview of CMIP5 and the experiment design. Bulletin of the American Meteorological Society, 93(4), 485–498. https://doi.org/10.1175/bams-d-11-00094.1
Wang, G., Hendon, H. H., Arblaster, J. M., Lim, E.-P., Abhik, S., & van Rensch, P. (2019). Compounding tropical and stratospheric forcing of the record low Antarctic sea-ice in 2016. Nature Communications, 10(1), 13. https://doi.org/10.1038/s41467-018-07689-7
Waugh, D. W., Banerjee, A., Fyfe, J. C., & Polvani, L. M. (2020). Contrasting recent trends in southern hemisphere Westerlies across different ocean basins. Geophysical Research Letters, 47, e2020GL088890. https://doi.org/10.1029/2020gl088890
Yang, D., Arblaster, J. M., Meehl, G. A., England, M. H., Lim, E.-P., Bates, S., & Rosenbloom, N. (2020). Role of tropical variability in driving decadal shifts in the Southern Hemisphere summertime eddy-driven jet. Journal of Climate, 33(13), 5445–5463. https://doi.org/10.1175/jcli-d-19-0604.1

Citing Literature

Volume126, Issue20

27 October 2021

e2021JD035023

The Role of Coupled Feedbacks in the Decadal Variability of the Southern Hemisphere Eddy-Driven Jet

Abstract

Key Points

Plain Language Summary

1 Introduction

2 Model and Experimental Design

3 Results