1. Introduction
The midlatitude atmospheric circulation is expected to alter under climate change (Shaw et al. 2016, and references therein). Such changes could have significant implications for the characteristics of extratropical cyclones (Tamarin-Brodsky and Kaspi 2017) and regional changes in extremes, such as droughts, heat waves, and wind storms (IPCC 2012; Li et al. 2018; Steptoe et al. 2018). However, there remain large uncertainties in the quantitative understanding of the midlatitude circulation response to changes in greenhouse gases (GHGs), aerosols, and natural climate forcings (Collins et al. 2013; Santer et al. 2013; Staten et al. 2012) and associated climate feedbacks such as through clouds and water vapor (Ceppi and Shepherd 2017; Voigt and Shaw 2015, 2016; Voigt et al. 2019). In the Southern Hemisphere (SH), reanalysis datasets and climate models show a strengthening and poleward shift of the midlatitude eddy-driven jet (EDJ) over recent decades (e.g., Polvani et al. 2011; Barnes and Polvani 2013), and a shift toward a more positive southern annular mode (SAM) index (Marshall 2003) with associated changes in storm tracks (e.g., Yin 2005; Chen and Held 2007; Bender et al. 2012), Rossby wave breaking (Ndarana et al. 2012), and atmospheric blocking (Dennison et al. 2016). The historical trends in SH circulation have been attributed to the effects of ozone depletion (primarily affecting the austral summer season) and rising greenhouse gases (Karpechko and Maycock et al. 2018 and references therein), with a potentially important role for internal climate variability (Garfinkel et al. 2015). Over the twenty-first century, general circulation models (GCMs) project further shifts in the SH midlatitude circulation associated with the effects of ozone recovery and scenario-dependent changes in GHGs (e.g., Ceppi and Hartmann 2013; Barnes and Polvani 2013; Barnes et al. 2014; Kushner et al. 2001; Previdi and Liepert 2007). While these major drivers are broadly known, the quantification of the importance of different forcings on the SH circulation remains uncertain. For example, there has been debate in the literature around the importance of anthropogenic aerosols for SH circulation trends (Gillett et al. 2013; Steptoe et al. 2016; Rotstayn 2013; Rotstayn et al. 2014; Choi et al. 2019).
The response of the eddy-driven jets to forcing is dependent on the spatial pattern of surface and atmospheric temperature changes (Butler et al. 2010; Ceppi and Shepherd 2017; Harvey et al. 2014, 2015; Murphy et al. 2002). Although the sign of the projected SH EDJ shift is generally robust across different climate models, they simulate a range of magnitudes of response to the same forcing scenario (Barnes and Polvani 2013; Barnes et al. 2014; Grise and Polvani 2014a). Grise and Polvani (2014a, 2016) showed that intermodel differences in the SH circulation response to an abrupt quadrupling in CO2 (4xCO2) can only be partly explained by differences in global mean surface air temperature (GSAT) responses and instead might be explained, to varying degrees depending on region and season, by increases in midlatitude and subtropical static stabilities (see also Butler et al. 2010) and by changes to surface and upper-tropospheric/lower-stratospheric equator-to-pole temperature gradients. Harvey et al. (2014) found that in CMIP5 models a significant fraction of the intermodel spread in projected SH storm track changes over the twenty-first century was linearly congruent with spread in upper- and lower-level temperature gradient responses. The radiative effects of clouds are also important for SH midlatitude circulation trends (Ceppi et al. 2014; Grise and Polvani 2014b; Albern et al. 2019). The EDJ response to forcing is likely to be seasonally varying due to the seasonal cycle in the background circulation (McGraw and Barnes 2016) and seasonally dependent factors such stratospheric polar vortex trends (Ceppi and Shepherd 2019). Given these various results, modeled changes in the midlatitude circulation therefore cannot be expected to simply scale with GSAT (Grise and Polvani 2014a, 2016, 2017; Ceppi et al. 2018), rendering efforts to constrain measures such as equilibrium climate sensitivity of limited value for projections of midlatitude circulation.
One emerging topic is the relative importance of “direct or rapid adjustment” and “indirect or slow feedback” processes for the midlatitude circulation response to forcing. Rapid atmospheric adjustments—which occur when a forcing is introduced but sea surface temperatures (SST) and sea ice are held fixed—are well known to affect the top of atmosphere energy budget through changes to clouds, thermal structure, and surface fluxes (Sherwood et al. 2015; Smith et al. 2018); however, the influence on large-scale circulation is underexplored. Grise and Polvani (2014c) decomposed the atmospheric circulation response to 4xCO2 in CMIP5 models into a rapid adjustment and an SST-driven component. While the rapid adjustment contributed a poleward shift of the EDJ, the overall response to 4xCO2 was dominated by SSTs, which are also responsible for the asymmetric midlatitude circulation response between hemispheres. Grise and Polvani (2017) further decomposed the SH circulation response to 4xCO2 by season and found that the midlatitude circulation shifts poleward on a similar time scale to GSAT change in austral summer (DJF), but adjusts more quickly than GSAT in austral winter (JJA). Ceppi et al. (2018) showed that nearly all of the SH EDJ response to 4xCO2 occurs within the first decade after the forcing is applied, despite less than 50% of the eventual quasi-equilibrium GSAT change occurring within this time. They also used a single GCM run in coupled and fixed-SST modes to separate rapid adjustments and the SST-mediated response. However, most of the above studies have primarily focused on the response to increasing CO2, so the relative importance of rapid adjustments and SST-driven feedbacks for the responses to other forcing agents is less well known. A recent study has shown that rapid adjustments to aerosol forcing contribute to changes in the location of the Hadley cell edge in the SH (Zhao et al. 2020), suggesting there may also be a role for the midlatitude circulation.
Idealized forcing studies allow a detailed examination of the circulation response to a single climate driver and an assessment of the differences in response across GCMs. However, a significant limitation is that most of the existing literature relies on the 4xCO2 experiment from CMIP5 and does not explore the responses to other major forcings such as non-CO2 GHGs, aerosols, and natural forcings. This study seeks to fill this gap by examining rapid adjustments and slow feedbacks in the SH midlatitude circulation in response to a number of idealized climate perturbations using output from the Precipitation Driver and Response Model Intercomparison Project (PDRMIP; Myhre et al. 2017).
2. Data and methods
a. Models and simulations
We use output from nine models participating in the PDRMIP (see Table S1 in the online supplemental material). Each model performed a control run using present-day conditions or preindustrial conditions (see Table S1), and five idealized abrupt single forcing experiments: 1) a doubling of carbon dioxide concentrations (2xCO2), 2) a tripling of methane concentrations (3xCH4), 3) a fivefold increase in sulfate aerosol concentrations or emissions (5xSO4), 4) a tenfold increase in black carbon aerosol concentrations or emissions (10xBC) and 5) a 2% increase in the solar constant (2%Sol). The distribution of the aerosol concentration or emissions perturbations is shown in Fig. 2 of Myhre et al. (2017). The scaling factors for the five idealized perturbations were chosen so that each experiment induced a similar magnitude of effective radiative forcing (ERF) to better enable climate responses to be compared across forcing agents (Smith et al. 2018). The perturbations are not designed to be based on a realistic future emissions scenario. In practice, the magnitudes of the multimodel mean (MMM) ERFs for 2xCO2, 5xSO4, and 2%Sol are comparable to within ~10%, whereas the ERFs for 3xCH4 and 10xBC are around 3 times smaller.
All experiments are performed in two model configurations: one set has sea surface temperatures and sea ice fixed to the control climatology (fSST) enabling the rapid adjustments to be diagnosed; the second set includes a coupled ocean and thus the full atmosphere–ocean response is captured. Note that the fSST experiments do not fix land temperatures, which may be important for some regional climate responses (Shaw and Voigt 2016). The ECHAM-HAM and CESM1-CAM4 models use a slab ocean and are excluded from the MMM and intermodel regression calculations. The fSST experiments are integrated for at least 15 years and the coupled experiments for at least 100 years (see Table S1). Responses are calculated by subtracting the time-mean climatology of the control simulation from the perturbed experiments. The fSST responses are diagnosed as the time mean excluding the first year, ensuring only the quasi-equilibrated adjusted state is captured (Richardson et al. 2016). In all cases the coupled response is diagnosed using the mean of years 50–100. This captures the quasi-equilibrated climate after the perturbations are applied, but neglects further climate responses, for example in the deep ocean, that will evolve for many hundreds of years before reaching a true equilibrium state (Caldeira and Myhrvold 2013).
The SPRINTARS and HadGEM2-ES models defined variables on pressure levels that are below the surface as missing data. Therefore, at some grid points over topography the MMM calculation excludes those models.
b. Midlatitude circulation response metrics
The response of the midlatitude circulation and its relation to the large-scale climate is diagnosed using metrics based on zonal winds and surface and air temperatures. The variables available from the different models to compute these diagnostics are listed in Table S2.
1) Latitude of EDJ
The speed and latitude of the EDJ (Umax and ϕEDJ) are diagnosed as the speed and location of the maximum zonal mean zonal wind
2) Upper- and lower-tropospheric temperature gradients
Following Harvey et al. (2014), diagnostics for upper (250 hPa) and lower (850 hPa) tropospheric temperature gradients (ΔT) are computed using the difference between the area-average temperature anomaly (T) in the tropics (30°S–30°N) and SH polar region (60°–90°S):
Previous studies using PDRMIP models that have focused on thermodynamic properties have often normalized metrics by GSAT or ERF to compare across the different perturbations (Richardson et al. 2019). However, since forced changes to the SH midlatitude circulation have been shown not to scale with GSAT change (e.g., Grise and Polvani 2017; Ceppi et al. 2018) we have chosen not to normalize the circulation changes in this study.
c. Statistical significance
Two main methods are used to assess the robustness of the model responses; both methods have been used to evaluate multimodel ensembles in the IPCC AR5 (IPCC 2013). The first method evaluates the agreement in the sign of the simulated anomalies across models. Hatching is applied where one or more models disagree on the sign of the anomaly. The second method quantifies the magnitude of the simulated anomalies compared to internal variability. To calculate each model’s internal variability in the coupled experiments, we calculate the standard deviation of consecutive 50-yr means sampled from the control runs and multiply this by the square root of 2 to reflect the uncertainty due to internal variability in the difference of two 50-yr means. Simulated changes are significant at the 95% level when the modulus is greater than 1.96σ, where σ is either taken from a given model or, in the case of the MMM, is the median standard deviation of the models.
For the fSST runs, a bootstrap was performed by selecting N years from the coupled control run, where N is the length of the fSST experiment in each model (see Table S1). The coupled control runs were used as they are longer than the fSST runs, enabling a more comprehensive assessment of internal variability to be attained. This was repeated 103 times to construct a probability density function. The standard deviation was then calculated and multiplied by the square root of 2 as before. Years are selected randomly as the fSST runs do not include any SST-mediated low-frequency variability. Lastly, intermodel regressions between variables are calculated using linear least squares regression.
3. Results
a. Surface temperature changes
Figures 1a–e show the MMM zonal mean SH near-surface temperature anomalies in the coupled experiments. All the forcings induce an increase in near-surface temperature except 5xSO4, which induces cooling as expected. The MMM global and annual mean surface air temperature (GSAT) changes in the 2xCO2, 3xCH4, 5xSO4, 10xBC, and 2%Sol experiments are 2.39, 0.70, −1.82, 0.85, and 2.38 K, respectively (see Table 1). The variation in magnitude of the MMM GSAT responses is primarily a consequence of differences in effective radiative forcing (ERF) between the perturbations (Richardson et al. 2019). The MMM ERFs for 2xCO2, 3xCH4, 5xSO4, 10xBC, and 2%Sol are 3.7, 1.1, −1.8, 1.3, and 4.1 W m−2, respectively (Tang et al. 2019). The GSAT changes also vary across models for the same experiment owing to differences in both ERF and modeled equilibrium climate sensitivity (Richardson et al. 2019). The zonal mean surface temperature anomalies show a heterogeneous pattern over the Southern Hemisphere, with a small amplification in the Antarctic and larger changes at latitudes with more landmass (Dong et al. 2009; Joshi et al. 2008). However, the changes in near-surface temperature gradient across the SH in the coupled experiments are consistently smaller than in the Northern Hemisphere where Arctic amplification leads to a weakened equator-to-pole temperature gradient in the lower troposphere (not shown).
Southern Hemisphere zonal mean near-surface temperature anomalies (K) in the five perturbation experiments (from top to bottom: 2xCO2, 3xCH4, 5xSO4, 10xBC, and 2%Sol) for (a)–(e) coupled experiments and (f)–(j) fixed SST experiments. The black line denotes the MMM, and gray shading indicates the model range. Solid lines indicate where the anomalies are significant at the 95% confidence level compared to internal variability (see section 2). Dashed lines show where the anomalies are not significant. Note that for each row, the y-axis scale in the right column is a factor of 5 smaller than in the left column.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
Multimodel mean differences in global mean surface temperature (K) for the five perturbation experiments in the coupled and fSST configurations. Model spread is shown in parentheses.
Figures 1f–j show the MMM zonal mean SH near-surface temperature anomalies in the fSST simulations. The anomalies are substantially smaller than in the coupled runs at all latitudes (<20%) but, as expected, the proportionately largest responses occur over land (e.g., Antarctica) where the surface temperatures are not fixed in the models. Over Antarctica, the aerosol perturbations cause small surface temperature changes of the opposite sign to those in the coupled experiments. At latitudes that are almost exclusively covered by ocean in the SH (~40°–60°S) the proportion of the temperature response that occurs in the fSST runs is very small (<10%), as expected. The GSAT anomalies in the fSST experiments are 0.28, 0.07, −0.09, 0.11, and 0.15 K (Table 1), which is around 5%–10% of the corresponding changes in the coupled experiments.
b. Zonal mean air temperature changes
Figure 2 shows the MMM annual and zonal mean air temperature anomalies in the five perturbation experiments; specifically, Figs. 2a–e show the coupled experiments, Figs. 2f–j show the fSST experiments, and Figs. 2k–o show the SST-mediated response (coupled minus fSST). As seen in Fig. 1, the amplitudes of the tropospheric temperature anomalies vary across the perturbations. The tropospheric response is strongest for 2xCO2 and 2%Sol and weakest for 3xCH4. The magnitude of the tropospheric temperature response to 5xSO4 lies between 2xCO2 and 3xCH4, but is opposite in sign. These differences can be understood from the ERF associated with each perturbation (see section 3a). Nevertheless, despite the different magnitudes, all the forcings except 10xBC induce a similar pattern of tropospheric temperature change in the coupled experiments, with larger anomalies in the tropical upper troposphere and near the Antarctic surface. This broad pattern of tropospheric temperature response is a robust feature in GCMs that accompanies changes in global mean surface temperature and associated climate feedbacks (IPCC 2013; Ceppi and Shepherd 2017).
Annual and multimodel mean Southern Hemisphere zonal mean temperature anomalies (K) in the five perturbation experiments (from left to right: 2xCO2, 3xCH4, 5xSO4, 10xBC, and 2%Sol): (a)–(e) the coupled experiments, (f)–(j) the fixed SST experiments, and (k)–(o) the coupled minus fixed SST differences. Stippling in the top and middle rows denotes where one or more models disagree on the sign of the local anomaly. Gray contours show the control climatology.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
The tropospheric temperature changes in 10xBC show a markedly different pattern to the other perturbations (Fig. 2d). Black carbon affects climate through absorption of shortwave radiation and so directly heats the atmosphere while inducing smaller surface temperature change (Stjern et al. 2017). The black carbon burden in the experiments is spatially dependent, with emissions in the SH primarily originating from biomass burning in central Africa, the Amazon, and Indonesia (Myhre et al. 2017).
Polar lower stratospheric temperature changes have been shown to be important for the tropospheric midlatitude circulation (e.g., Harvey et al. 2014; Thompson and Solomon 2002; Grise and Polvani 2017). In the polar lower stratosphere, all perturbations except 10xBC induce temperature changes in the coupled experiment that are opposite in sign to the tropospheric response. The 2xCO2 experiment induces strong radiative cooling in the stratosphere that increases with height (Fels et al. 1980). The 10xBC perturbation induces a weak warming in the Antarctic lower stratosphere.
The fSST temperature anomalies due to the five perturbations show markedly different patterns from the coupled experiments (Figs. 2f–j). The enhanced tropical upper tropospheric temperature changes found for 2xCO2, 3xCH4, 5xSO4, and 2%Sol are largely absent in the fSST experiments, indicating this structure is predominantly due to SST-driven feedbacks through moist convective adjustment in the tropics (Figs. 2k–o). Instead the tropospheric temperature changes in the 2xCO2, 3xCH4, 5xSO4, and 2%Sol fSST experiments are weaker and more homogeneous. 2xCO2 is the only perturbation that induces strong radiative cooling in the stratosphere in the fSST experiments (Smith et al. 2018). There is also a stratospheric heating response in 2%Sol, which is mainly due to shortwave absorption by ozone. In contrast to 2xCO2, 3xCH4, 5xSO4, and 2%Sol, the fSST temperature response in 10xBC more closely resembles the coupled experiment showing that the direct shortwave absorption by black carbon dominates the anomalous atmospheric heating. The difference between the coupled and fSST temperature anomalies for 10xBC reveals a pattern that resembles the response to the other forcings (Figs. 2k–o), with a magnitude similar to the coupled 3xCH4 experiment. This is consistent with the similar magnitudes of ERF and GSAT change for the 10xBC and 3xCH4 perturbations (Table 1). The overall temperature response in the coupled 10xBC experiment is therefore relatively less affected by SST-driven feedbacks and is predominantly due to shortwave absorption.
The results in Figs. 1 and 2 demonstrate the uneven distribution of surface and atmospheric temperature changes induced by the different climate drivers, as well as the important role of SSTs in determining these changes particularly for 2xCO2, 3xCH4, 5xSO4, and 2%Sol. The coincident responses of the midlatitude circulation are investigated in the following sections.
c. Zonal wind changes
1) Austral summer (DJF)
The MMM
As in Fig. 2, but for the
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
All the fSST perturbation experiments (Figs. 3f–j) except 3xCH4 also show a dipole pattern of tropospheric
Figure 4 shows longitude–latitude SH zonal wind anomalies at 850 hPa (U850) in the five perturbation experiments. The left and right columns in Fig. 4 show the coupled and fSST experiments, respectively. For all perturbations there is a high degree of longitudinal symmetry in the midlatitude zonal wind responses in the coupled experiments. This is consistent with there being a single well-defined EDJ in austral summer associated with a circumpolar storm track (Nakamura and Shimpo 2004). There is less consistency across the models in the coupled response to 10xBC particularly between 60°E and 180°, which could be partially explained by the differences in black carbon burden in the different models (Stjern et al. 2017). The models agree on the sign of the U850 anomalies in the region of the EDJ in the other coupled forcing experiments.
Multimodel mean Southern Hemisphere DJF 850-hPa zonal wind anomalies (m s−1) in the five perturbation experiments (from top to bottom: 2xCO2, 3xCH4, 5xSO4, 10xBC, and 2%Sol) for (a)–(e) coupled experiments and (f)–(j) fixed SST experiments. Stippling is as in Fig. 2. Gray contours show the control climatology.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
The fSST experiments (Figs. 4f–j) also show a high degree of longitudinal symmetry in the midlatitude U850 changes. Although the responses are smaller, and the agreement across the models lower than in the coupled experiments, the MMM pattern of U850 anomalies in the fSST experiments strongly resembles the coupled responses particularly in the midlatitudes. The rapid adjustment is proportionately largest for 10xBC, where in the southern Indian Ocean sector and south of Australia the fSST U850 anomalies are comparable in magnitude to the fully coupled response and up to around 60% of the magnitude near South America. For 2xCO2, 5xSO4, and 2%Sol the magnitude of the U850 anomalies in the fSST experiments is typically around 30%–60% of that in the coupled experiments in the important EDJ region where the response is most robust.
The individual model wind anomalies can be seen in Fig. 5, which shows zonal mean U850 (
DJF 850-hPa
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
In summary, this section has shown the rapid adjustments of the SH midlatitude circulation in DJF resemble a weaker version of the coupled responses.
2) Austral winter (JJA)
Figure 6 shows the MMM
As in Fig. 2, but for the
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
A weaker midlatitude zonal wind response to forcing in JJA, as compared to DJF, has been previously identified in response to increased CO2 (e.g., Grise and Polvani 2016) and stratospheric water vapor changes (Maycock et al. 2013) and in idealized models with applied thermal forcings (McGraw and Barnes 2016). Despite the smaller magnitude, there is high agreement across the models on the sign of the zonal wind changes in the region of the EDJ.
The fSST experiments show
Figure 7 shows the U850 anomalies in JJA. The U850 responses are generally less longitudinally symmetric than in DJF (cf. Fig. 4), with the westerly anomalies (easterlies for 5xSO4) being more equatorward in the Indian and Pacific Ocean sectors and more poleward in the Atlantic sector. The fSST responses in JJA (Figs. 7f–j) show that the 2xCO2 and 3xCH4 perturbations induce the strongest U850 response in the southern Indian Ocean and South Pacific sectors and a comparatively weaker response in the South Atlantic. The rapid adjustment to 10xBC is proportionately stronger than for the other forcings in both JJA and DJF. Interestingly, to the south of Australia in the 10xBC fSST experiment the response is opposite in sign to the coupled anomalies. This indicates that the SST-mediated changes must override the direct atmospheric response in this region. As seen in Fig. 6j, in contrast to the other perturbations the U850 anomalies in the 2%Sol fSST experiment are weak and not consistent in sign across the models.
As in Fig. 4, but for the U850 differences (m s−1) in JJA.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
Figure 8 shows the latitudinal profiles of
As in Fig. 5, but for the 850-hPa
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
The U850 responses in the fSST experiments are less consistent across the models, with some responding with the opposite sign such as IPSL-CM5A in the 3xCH4 experiment and SPRINTARS in the 5xSO4 and 2%Sol experiments. On average, the fSST U850 anomalies in JJA are smaller in magnitude and less consistent in sign across the models than in DJF.
d. Jet latitude and speed
This section puts the SH zonal wind changes shown in section 3c into the context of changes in the latitude and speed of the EDJ. Figure 9 shows the EDJ latitude and speed anomalies in DJF for the five perturbation experiments for the individual models and the MMM. Note that SPRINTARS simulates a broad jet without a well-defined maximum and hence there is a large error on the estimate of the EDJ latitude so it is excluded from the MMM calculation.
Southern Hemisphere EDJ latitude (Δϕ) and speed (ΔUmax) anomalies in the five perturbation experiments in DJF. Coupled and fixed SST EDJ latitude shift anomalies are shown in red and pink (left y axis), respectively, and coupled and fixed SST EDJ speed anomalies are in blue and turquoise (right y axis), respectively. Filled symbols denote changes that are significant at the 95% confidence level compared to internal variability (see section 2). The MMM is denoted by the solid line, which is black where the MMM change is significant at the 95% confidence level compared to the median internal variability or gray otherwise. Slab ocean models and SPRINTARS that are not included in the MMM calculation are shown in gray symbols.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
Figure 9 shows a consistent poleward shift and strengthening of the EDJ across the models in the DJF season for the 2xCO2 and 2%Sol experiments, and the consistent equatorward shift and weakening of the EDJ in the 5xSO4 experiment. In the MMM, the EDJ shifts poleward by ~2° and strengthens by ~0.7 m s−1 in the coupled 2xCO2 and 2%Sol experiments. There is an equal and opposite MMM response of the jet speed in 5xSO4, but a smaller equatorward jet shift of ~1.5°. The MMM changes in 10xBC show a poleward shift and increase in jet speed, which is around half the magnitude of 2xCO2 and 2%Sol (1.4° and 0.3 m s−1, respectively). In all the coupled experiments, the sign of the jet shift in DJF is consistent across the models and the MMM response is significant compared to internal variability. The same is true for jet speed except for the 3xCH4 experiment, which produces a nonsignificant MMM change, and for 10xBC where two models simulate a reduction in jet speed.
As in previous sections, the rapid adjustment in DJF is proportionately largest for 10xBC, with a magnitude of ~75% and ~90% for the MMM jet shift and jet speed changes, respectively, compared to the coupled experiment. In the other experiments, the magnitude of the rapid adjustment is a smaller yet substantial fraction of the overall coupled responses (typically 20%–30% of the jet shift and 20%–35% of jet speed). Furthermore, across all models and all experiments there is a significant correlation between the jet shift (ΔφEDJ) in the fSST and coupled experiments with a pooled R2[ΔφfSST(ALL), Δφcoupled(ALL)]DJF = 0.30 (P = 0.003). For the jet speed (ΔUmax) anomalies the pooled R2[ΔUfSST(ALL), ΔUcoupled(ALL)]DJF = 0.37 (P < 0.001). This suggests that around one-third of the intermodel spread in the DJF jet shift and jet speed responses to the perturbations can be interpreted as being due to spread in the rapid adjustment.
Figure 10 shows the EDJ latitude and speed anomalies for JJA. There is a less consistent picture of the responses to the perturbations than was found in DJF. Nevertheless, consistent with the results in section 3c, the MMM responses in JJA tend to project more strongly onto a change in EDJ strength, with larger anomalies than in DJF, while there are smaller changes in EDJ latitude compared to DJF by around a factor of 2. Interestingly, the coupled jet shift to 5xSO4 is opposite in sign and significant in JJA and the change in jet speed is smaller than in DJF, in contrast to the other experiments. The rapid adjustment accounts for a larger proportion of the overall coupled changes in JJA than in DJF in the 2xCO2 experiment. In contrast to DJF, there is a weak and nonsignificant relationship between the jet shift and jet speed anomalies in the fSST and coupled experiments in JJA. This suggests that in JJA the intermodel spread in responses in the coupled simulations is predominantly driven by other factors than rapid adjustments.
e. Relationship to temperature gradients
In the extratropics, the meridional temperature gradient is proportional to the vertical wind shear through thermal wind balance. Harvey et al. (2014) found that in CMIP5 models differences in upper- and lower-level temperature gradients could explain a significant fraction of the intermodel spread in projected SH storm track changes over the twenty-first century. Here we examine the changes to temperature gradients in the PDRMIP experiments to determine the differences in the rapid adjustment and coupled zonal wind responses discussed above.
Figure 11 shows the annual mean upper- (250 hPa) and lower- (850 hPa) tropospheric meridional temperature gradient anomalies ΔT for the five climate perturbations in the coupled and fSST experiments. A negative (positive) anomaly shows a reduction (increase) in the tropics-to-pole temperature gradient. There were found to be only small seasonal differences (see Figs. S1 and S2), partly because polar amplification is smaller in the SH than in the Arctic, so for simplicity the changes in annual mean temperature gradients are shown.
Annual mean 850- and 250-hPa meridional temperature gradient changes (30°N–30°S minus 60°–90°S) in the five perturbation experiments. The ΔT850 coupled and fixed SST anomalies are shown in red and pink (left y axis), respectively, and ΔT250 coupled and fixed SST anomalies are in blue and turquoise (right y axis), respectively. Filled symbols denote changes that are significant at the 95% confidence level compared to internal variability (see section 2). The MMM is denoted by the solid line, which is black where the MMM change is significant at the 95% confidence level compared to the median internal variability or gray otherwise. Slab ocean models that are not included in the MMM calculation are shown in gray symbols.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
The models agree on the sign of ΔT250 in all the coupled experiments. The few models that show a negative ΔT850 for some perturbations are those which produce larger polar amplification (CESM1-CAM4, CESM1-CAM5). CESM1-CAM4 is a consistent outlier, which may be related to it using a simplified slab ocean model. As was done for the anomalies in jet latitude and jet speed, the slab ocean models have been excluded from the MMM. All the coupled experiments show a greater increase in ΔT250 than in ΔT850, predominantly due to the large tropical upper tropospheric temperature changes [r(ΔT250, T250,trop) > 0.87 for all experiments; see Table S3]. Similar differences between ΔT850 and ΔT250 in the SH were shown by Harvey et al. (2014) using CMIP5 models.
For all the experiments except 10xBC, the rapid adjustment in ΔT850 and ΔT250 is comparatively small, which is a consequence of the direct tropospheric heating induced by the forcings being weak and fairly uniform across the SH (Fig. 2). SST feedbacks therefore dominate the overall changes in ΔT850 and ΔT250 for those perturbations. While there are larger temperature changes in the troposphere and lower stratosphere in the 10xBC fSST experiment (Fig. 2), in the tropics these changes peak in the lower stratosphere above 250 hPa and it is only in the subtropics that larger temperature anomalies penetrate into the upper troposphere. Hence, the MMM rapid adjustment in ΔT250 is near-zero for 10xBC. There are two outliers, however, with CCCma showing a strong negative ΔT250 and HadGEM2 showing a positive ΔT250. These models have the largest black carbon burden of all the PDRMIP models (Stjern et al. 2017) and are both driven by aerosol emissions rather than concentrations. This suggests there are differences in the distribution of black carbon and hence tropospheric heating between the two models. An inspection of the individual model zonal mean air temperature responses reveals that CCCma produces strong upper-tropospheric and stratospheric warming in a pattern that resembles the MMM (not shown). However, the zonal mean response to 10xBC in HadGEM2 more closely resembles the response to 2xCO2, producing cooling in the stratosphere. This cooling may be caused by the substantial increase in stratospheric water vapor in the HadGEM2 model. Except for 10xBC, the intermodel spread in both ΔT250 and ΔT850 is consistently larger in the coupled experiments than for fSST, reflecting the increase in model spread due to differences in climate feedbacks. For most of the perturbations, the different responses of SSTs are therefore the dominant cause of the intermodel spread in ΔT250 and ΔT850.
In the coupled experiments, ΔT250 and ΔT850 are significantly correlated across all models and experiments in both DJF and JJA [r(ΔT250, ΔT850)DJF = 0.87 (P < 0.001); r(ΔT250, ΔT850)JJA = 0.75 (P < 0.001)]. This is higher than the correlation between ΔT250 and ΔT850 in the SH across CMIP5 models (0.45 and 0.50 for DJF and JJA, respectively; Harvey et al. 2014). The correlation between ΔT250 and ΔT850 means that a regression of zonal wind responses onto either index yields qualitatively similar results. We proceed by examining the relationship between the zonal wind responses and ΔT850 across the models, with the caveat that in these experiments using this method we cannot distinguish the relative importance of changes to upper- and lower-level temperature gradients.
Figure 12 shows the fraction of intermodel variance in U850 response that is explained by ΔT850, defined as the coefficient of determination (R2). The fraction of variance explained is up to 80%–90% in the EDJ region in all of the coupled experiments except 3xCH4. When a regression is performed across all coupled models and all perturbation experiments (Fig. 12f), a significant relationship is found across the entire midlatitude region, with around 50%–90% of the variance in local midlatitude U850 responses explained by ΔT850. This indicates that the regression onto ΔT850 is a useful measure for capturing much of the model spread in midlatitude circulation responses (cf. Harvey et al. 2014).
The fraction of intermodel variance in ΔU850 explained by the regression onto ΔT850, defined as the coefficient of determination (R2). Stippling indicates where the correlation is statistically significant (P < 0.05).
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
In general the perturbations that generate the largest changes in temperature gradients (2xCO2 and 2%Sol) also show the largest zonal wind responses (Figs. 2 and 3), as expected from thermal wind balance. To determine if there are different intermodel relationships between the U850 and ΔT850 responses for the perturbations, Fig. 13 shows the regression of U850 against ΔT850 for the five perturbations. As the strongest and most consistent responses of the EDJ to the forcings occur in austral summer, we focus on DJF.
Intermodel regression of DJF U850 vs DJF ΔT850 responses (m s−1 K−1) for (a)–(e),(g)–(k) the five perturbation experiments and (f),(l) across all experiments: (left) coupled experiments and (right) fixed SST experiments. Stippling is as in Fig. 12.
Citation: Journal of Climate 33, 22; 10.1175/JCLI-D-19-1015.1
The spatial pattern of the intermodel regression between U850 and ΔT850 responses in the coupled experiments is similar across all forcings despite the different magnitudes of responses and spatial patterns shown earlier. The patterns resemble those shown in Fig. 4 with an enhanced low-level temperature gradient (positive ΔT850) being associated with a dipole U850 anomaly pattern with stronger westerlies on the poleward flank of the jet and weaker westerlies on the equatorward flank. While the magnitudes of this pattern are largely similar across the five perturbations, there is a suggestion of a weaker regression coefficient in the 5xSO4 experiment, which could be due to the hemispheric asymmetry of the forcing (Richardson et al. 2019). There may also be a difference in the sensitivity of the zonal wind response to tropospheric cooling, as in 5xSO4, as compared to the warming produced by the other forcings.
The intermodel regression applied to the fSST experiments reveals a similar dipole pattern of midlatitude U850 anomalies (Figs. 13g–k); however, the magnitude of the regression slope is around double that found in the coupled experiments. This means that the U850 responses in the fSST experiments are proportionately larger than would be predicted by ΔT850 using the relationship derived from the coupled experiments. While the rapid adjustment accounts for a minor part of the total coupled response for most of the perturbations (except 10xBC), it is proportionately larger than expected from the associated differences in ΔT850. This is interesting and suggests that the processes captured within the rapid adjustment, such as changes to clouds, may affect the midlatitude circulation in a manner not captured by a simple measure of large-scale temperature gradient change. Broadly similar differences between the coupled and fSST regressions is seen in JJA (see Fig. S3). A regression of SH zonal mean air temperature onto ΔT850 in the coupled simulations reveals a tropospheric pattern that is dominated by warming in the tropics and midlatitudes (Figs. S4a–e). However, in the fSST simulations, the pattern of temperature anomalies regressed onto ΔT850 is markedly different, with a warming (cooling) dipole in the lower troposphere in the middle (high) latitudes (Figs. S4f–j). This hints toward a possible reason for the different relationships of U850 anomalies to ΔT850 in the fSST and coupled experiments.
4. Conclusions
This study has analyzed the rapid adjustment (i.e., the meteorological response to forcing in the absence of SST and sea ice changes) and coupled atmosphere–ocean response of the Southern Hemisphere (SH) midlatitude circulation to five idealized climate perturbations: 2x carbon dioxide, 3x methane, 5x sulfate aerosol, 10x black carbon aerosol, and +2% solar constant. Previous studies have predominantly focused on rapid adjustments to increased CO2 (e.g., Grise and Polvani 2014a,b, 2016, 2017; Ceppi and Shepherd 2017), so the novelty of this study lies in its consideration of a broader set of climate perturbations applied in nine global climate models. The rapid adjustment occurs as a result of atmospheric and land surface heating, including changes to latent heating and surface fluxes, while the coupled response is mediated by SST changes.
In the coupled experiments, the response of the SH circulation projects more strongly onto a shift in the EDJ in austral summer, whereas in winter the response projects more strongly onto changes in jet strength. This is similar to other studies examining the seasonality of the SH midlatitude circulation response to a constant year-round forcing (e.g., Maycock et al. 2013; McGraw and Barnes 2016). These seasonal differences could be explained by differences in the barotropic responses between seasons, with a stronger upper-level barotropic response in summer, including changes to horizontal wave propagation and breaking (Nie et al. 2016). Barotropic Rossby wave breaking transports eddy momentum and so affects the position and speed of the EDJ (Robinson 2006). In DJF, the SH EDJ is more zonally symmetric, narrower, and stronger than in JJA, when the climatological westerlies extend farther toward the subtropics, especially in the Pacific sector (Ceppi and Hartmann 2013; Nakamura and Shimpo 2004). Our results corroborate this, with more regional heterogeneity in the low-level zonal wind responses in JJA than in DJF. The results highlight the need to consider the seasonality of the midlatitude circulation response to forcing.
In DJF, the magnitude of the EDJ shift due to the rapid adjustment is ~20%–30% of the coupled response for most of the perturbations except 10xBC where the rapid adjustment is 75% of the coupled jet shift. Hence, SST feedbacks exert an important control on the midlatitude circulation response to forcings that induce substantial global surface temperature changes, but the rapid adjustment is still a substantial fraction of the overall response. While the rapid adjustment in jet latitude in DJF is statistically significant in some models, the magnitude of the multimodel mean is small compared to internal variability for all perturbations than BCx10, so it is very unlikely that rapid adjustments in the SH midlatitude circulation could be observed. Nevertheless, for all experiments there is strong consistency in the sign of the jet latitude and speed anomalies across models, which is unlikely to be a result of internal variability. The proportionately stronger rapid adjustment to 10xBC indicates that the response is less coupled to surface temperature changes and is strongly influenced by absorption of shortwave radiation by black carbon. The magnitude of the rapid adjustment to 2xCO2 is similar to previous studies (e.g., Grise and Polvani 2014c) and also shows the rapid adjustment is comparable for methane, sulfate (although opposite in sign), and solar forcing, despite them acting through different radiative mechanisms (e.g., the extent of stratospheric cooling/heating). Despite the aerosol forcing being concentrated in the Northern Hemisphere (Richardson et al. 2019), the perturbations still induce detectable changes in the SH circulation in both the rapid adjustment and coupled experiments. Integrated assessment models simulate substantial changes in emissions of aerosols such as black carbon and sulfate over the coming decades. These results suggest this could impact the large-scale circulation in the Southern Hemisphere, though there remains debate around the role of aerosol forcing for historical observed SH climate (Gillett et al. 2013; Steptoe et al. 2018; Choi et al. 2019).
A variance analysis shows that up to 80%–90% of the intermodel spread in local midlatitude 850-hPa zonal wind responses (U850) in DJF is linearly congruent with the spread in the 850-hPa tropics-to-pole temperature gradient anomaly (ΔT850), with comparable spatial patterns for all five perturbations. However, the gradient of the relationship between U850 and ΔT850 is around a factor of 2 larger for the rapid adjustment than the equivalent coupled experiment, so the magnitude of the rapid adjustment in circulation is proportionately larger than would be predicted from the U850–ΔT850 relationship in the coupled experiments. This is interesting and suggests that the processes captured within the rapid adjustment, such as land temperature changes and clouds, are affecting the midlatitude circulation in a manner that cannot be quantitatively captured by a simple measure of large-scale temperature gradient change. This also demonstrates why common global climate change measures, such as GSAT, are generally not useful measures for determining regional midlatitude climate change (Grise and Polvani 2016; Ceppi et al. 2018). The results demonstrate the need to understand both rapid adjustments and SST-mediated feedbacks to fully understand the SH midlatitude circulation response to forcing.
Acknowledgments
TW was funded by a PGR studentship from the NERC SPHERES Doctoral Training Partnership (NE/L002574/1). ACM was funded by a NERC Independent Research Fellowship (NE/M018199/1). ACM and PMF were funded by the European Union’s Horizon 2020 Research and Innovation Programme under Grant Agreement 820829 (CONSTRAIN) and UKRI NERC Grant NE/N006038/1 (SMURPHS). T. T. was supported by the supercomputer system of the National Institute for Environmental Studies, Japan, and JSPS KAKENHI Grant JP19H05669. D. O. and A. K. were supported by the Norwegian Research Council through the projects EVA (229771), EarthClim (207711/E10), NOTUR (nn2345k), and NorStore (ns2345k). T. A. was supported by the Joint U.K. BEIS/Defra Met Office Hadley Centre Climate Programme (GA01101). B. H. S. and G. M. were funded by the Research Council of Norway, through the grant NAPEX (229778). O. B. acknowledges HPC resources from TGCC under the gencmip6 allocation provided by GENCI (Grand Equipement National de Calcul Intensif). The authors thank Paulo Ceppi and one anonymous reviewer for their constructive comments that improved the article.
Data availability statement
The PDRMIP model output is publicly available (for data access, visit http://www.cicero.uio.no/en/PDRMIP/PDRMIP-data-access).
REFERENCES
Albern, N., A. Voigt, and J. G. Pinto, 2019: Cloud-radiative impact on the regional responses of the midlatitude jet streams and storm tracks to global warming. J. Adv. Model. Earth Syst., 11, 1940–1958, https://doi.org/10.1029/2018MS001592.
Barnes, E. A., and L. Polvani, 2013: Response of the midlatitude jets, and of their variability, to increased greenhouse gases in the CMIP5 models. J. Climate, 26, 7117–7135, https://doi.org/10.1175/JCLI-D-12-00536.1.
Barnes, E. A., N. W. Barnes, and L. M. Polvani, 2014: Delayed Southern Hemisphere climate change induced by stratospheric ozone recovery, as projected by the CMIP5 models. J. Climate, 27, 852–867, https://doi.org/10.1175/JCLI-D-13-00246.1.
Bender, F. A. M., V. Ramanathan, and G. Tselioudis, 2012: Changes in extratropical storm track cloudiness 1983–2008: Observational support for a poleward shift. Climate Dyn., 38, 2037–2053, https://doi.org/10.1007/s00382-011-1065-6.
Butler, A. H., D. W. J. Thompson, and R. Heikes, 2010: The steady-state atmospheric circulation response to climate change–like thermal forcings in a simple general circulation model. J. Climate, 23, 3474–3496, https://doi.org/10.1175/2010JCLI3228.1.
Caldeira, K., and N. P. Myhrvold, 2013: Projections of the pace of warming following an abrupt increase in atmospheric carbon dioxide concentration. Environ. Res. Lett., 8, 034039, https://doi.org/10.1088/1748-9326/8/3/034039.
Ceppi, P., and D. L. Hartmann, 2013: On the speed of the eddy-driven jet and the width of the Hadley cell in the Southern Hemisphere. J. Climate, 26, 3450–3465, https://doi.org/10.1175/JCLI-D-12-00414.1.
Ceppi, P., and T. G. Shepherd, 2017: Contributions of climate feedbacks to changes in atmospheric circulation. J. Climate, 30, 9097–9118, https://doi.org/10.1175/JCLI-D-17-0189.1.
Ceppi, P., and T. G. Shepherd, 2019: The role of the stratospheric polar vortex for the austral jet response to greenhouse gas forcing. Geophys. Res. Lett., 46, 6972–6979, https://doi.org/10.1029/2019GL082883.
Ceppi, P., M. D. Zelinka, and D. L. Hartmann, 2014: The response of the Southern Hemispheric eddy-driven jet to future changes in shortwave radiation in CMIP5. Geophys. Res. Lett., 41, 3244–3250, https://doi.org/10.1002/2014GL060043.
Ceppi, P., G. Zappa, T. G. Shepherd, and J. M. Gregory, 2018: Fast and slow components of the extratropical atmospheric circulation response to CO2 forcing. J. Climate, 31, 1091–1105, https://doi.org/10.1175/JCLI-D-17-0323.1.
Chen, G., and I. M. Held, 2007: Phase speed spectra and the recent poleward shift of Southern Hemisphere surface westerlies. Geophys. Res. Lett., 34, L21805, https://doi.org/10.1029/2007GL031200.
Choi, J., S. W. Son, and R. J. Park, 2019: Aerosol versus greenhouse gas impacts on Southern Hemisphere general circulation changes. Climate Dyn., 52, 4127–4142, https://doi.org/10.1007/s00382-018-4370-5.
Collins, M., and Coauthors, 2013: Long-term climate change: Projections, commitments and irreversibility. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1029–1136.
Dennison, F. W., A. McDonald, and O. Morgenstern, 2016: The influence of ozone forcing on blocking in the Southern Hemisphere. J. Geophys. Res. Atmos., 121, 14 358–14 371, https://doi.org/10.1002/2016JD025033.
Dong, B., J. M. Gregory, and R. T. Sutton, 2009: Understanding land–sea warming contrast in response to increasing greenhouse gases. Part I: Transient adjustment. J. Climate, 22, 3079–3097, https://doi.org/10.1175/2009JCLI2652.1.
Fels, S. B., J. D. Mahlman, M. D. Schwarzkopf, and R. W. Sinclair, 1980: Stratospheric sensitivity to perturbations in ozone and carbon dioxide: Radiative and dynamical response. J. Atmos. Sci., 37, 2265–2297, https://doi.org/10.1175/1520-0469(1980)037<2265:SSTPIO>2.0.CO;2.
Garfinkel, C. I., D. W. Waugh, and L. M. Polvani, 2015: Recent Hadley cell expansion: The role of internal atmospheric variability in reconciling modeled and observed trends. Geophys. Res. Lett., 42, 10 824–10 831, https://doi.org/10.1002/2015GL066942.
Gillett, N. P., J. C. Fyfe, and D. E. Parker, 2013: Attribution of observed sea level pressure trends to greenhouse gas, aerosol, and ozone changes. Geophys. Res. Lett., 40, 2302–2306, https://doi.org/10.1002/grl.50500.
Grise, K. M., and L. M. Polvani, 2014a: Is climate sensitivity related to dynamical sensitivity? A Southern Hemisphere perspective. Geophys. Res. Lett., 41, 534–540, https://doi.org/10.1002/2013GL058466.
Grise, K. M., and L. M. Polvani, 2014b: Southern Hemisphere cloud-dynamics biases in CMIP5 models and their implications for climate projections. J. Climate, 27, 6074–6092, https://doi.org/10.1175/JCLI-D-14-00113.1.
Grise, K. M., and L. M. Polvani, 2014c: The response of midlatitude jets to increased CO2: Distinguishing the roles of sea surface temperature and direct radiative forcing. Geophys. Res. Lett., 41, 6863–6871, https://doi.org/10.1002/2014GL061638.
Grise, K. M., and L. M. Polvani, 2016: Is climate sensitivity related to dynamical sensitivity? J. Geophys. Res. Atmos., 121, 5159–5176, https://doi.org/10.1002/2015JD024687.
Grise, K. M., and L. M. Polvani, 2017: Understanding the time scales of the tropospheric circulation response to abrupt CO2 forcing in the Southern Hemisphere: Seasonality and the role of the stratosphere. J. Climate, 30, 8497–8515, https://doi.org/10.1175/JCLI-D-16-0849.1.
Harvey, B. J., L. C. Shaffrey, and T. J. Woollings, 2014: Equator-to-pole temperature differences and the extra-tropical storm track responses of the CMIP5 climate models. Climate Dyn., 43, 1171–1182, https://doi.org/10.1007/s00382-013-1883-9.
Harvey, B. J., L. C. Shaffrey, and T. J. Woollings, 2015: Deconstructing the climate change response of the Northern Hemisphere wintertime storm tracks. Climate Dyn., 45, 2847–2860, https://doi.org/10.1007/s00382-015-2510-8.
IPCC, 2012: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. C. B. Field et al., Eds., Cambridge University Press, 582 pp.
IPCC, 2013: Climate Change 2013: The Physical Science Basis. T. F. Stocker et al., Eds., Cambridge University Press, 1535 pp.
Joshi, M. M., J. M. Gregory, M. J. Webb, D. M. H. Sexton, and T. C. Johns, 2008: Mechanisms for the land/sea warming contrast exhibited by simulations of climate change. Climate Dyn., 30, 455–465, https://doi.org/10.1007/s00382-007-0306-1.
Karpechko, A. Y., A. C. Maycock, M. Abalos, H. Akiyoshi, J. M. Arblaster, C. I. Garfinkel, K. H. Rosenhof, and M. Sigmond, 2018: Stratospheric ozone changes and climate. Scientific Assessment of Ozone Depletion: 2018, Global Ozone Research and Monitoring Project Rep. 58, World Meteorological Organization, 44–46.
Kushner, P. J., I. M. Held, and T. L. Delworth, 2001: Southern Hemisphere atmospheric circulation response to global warming. J. Climate, 14, 2238–2249, https://doi.org/10.1175/1520-0442(2001)014<0001:SHACRT>2.0.CO;2.
Li, C., and Coauthors, 2018: Midlatitude atmospheric circulation responses under 1.5 and 2.0°C warming and implications for regional impacts. Earth Syst. Dyn., 9, 359–382, https://doi.org/10.5194/esd-9-359-2018.
Marshall, G. J., 2003: Trends in the southern annular mode from observations and reanalyses. J. Climate, 16, 4134–4143, https://doi.org/10.1175/1520-0442(2003)016<4134:TITSAM>2.0.CO;2.
Maycock, A. C., M. M. Joshi, K. P. Shine, and A. A. Scaife, 2013: The circulation response to idealized changes in stratospheric water vapor. J. Climate, 26, 545–561, https://doi.org/10.1175/JCLI-D-12-00155.1.
McGraw, M. C., and E. A. Barnes, 2016: Seasonal sensitivity of the eddy-driven jet to tropospheric heating in an idealized AGCM. J. Climate, 29, 5223–5240, https://doi.org/10.1175/JCLI-D-15-0723.1.
Murphy, B. F., P. Pettré, and I. Simmonds, 2002: Effects of changing baroclinicity on the Southern Hemisphere extratropical circulation. Quart. J. Roy. Meteor. Soc., 128, 1807–1826, https://doi.org/10.1256/003590002320603421.
Myhre, G., and Coauthors, 2017: PDRMIP: A precipitation driver and response model intercomparison project—Protocol and preliminary results. Bull. Amer. Meteor. Soc., 98, 1185–1198, https://doi.org/10.1175/BAMS-D-16-0019.1.
Nakamura, H., and A. Shimpo, 2004: Seasonal variations in the Southern Hemisphere storm tracks and jet streams as revealed in a reanalysis dataset. J. Climate, 17, 1828–1844, https://doi.org/10.1175/1520-0442(2004)017<1828:SVITSH>2.0.CO;2.
Ndarana, T., D. W. Waugh, L. M. Polvani, G. J. P. Correa, and E. P. Gerber, 2012: Antarctic ozone depletion and trends in tropopause Rossby wave breaking. Atmos. Sci. Lett., 13, 164–168, https://doi.org/10.1002/asl.384.
Nie, Y., Y. Zhang, G. Chen, and X.-Q. Yang, 2016: Delineating the barotropic and baroclinic mechanisms in the midlatitude eddy-driven jet response to lower-tropospheric thermal forcing. J. Atmos. Sci., 73, 429–448, https://doi.org/10.1175/JAS-D-15-0090.1.
Polvani, L. M., D. W. Waugh, G. J. P. Correa, and S. W. Son, 2011: Stratospheric ozone depletion: The main driver of twentieth-century atmospheric circulation changes in the Southern Hemisphere. J. Climate, 24, 795–812, https://doi.org/10.1175/2010JCLI3772.1.
Previdi, M., and B. G. Liepert, 2007: Annular modes and Hadley cell expansion under global warming. Geophys. Res. Lett., 34, L22701, https://doi.org/10.1029/2007GL031243.
Richardson, T. B., B. H. Samset, T. Andrews, G. Myhre, and P. M. Forster, 2016: An assessment of precipitation adjustment and feedback computation methods. J. Geophys. Res. Atmos., 121, 11 608–11 619, https://doi.org/10.1002/2016JD025625.
Richardson, T. B., and Coauthors, 2019: Efficacy of climate forcings in PDRMIP models. J. Geophys. Res. Atmos., 124, 12 824–12 844, https://doi.org/10.1029/2019JD030581.
Robinson, W. A., 2006: On the self-maintenance of midlatitude jets. J. Atmos. Sci., 63, 2109–2122, https://doi.org/10.1175/JAS3732.1.
Rotstayn, L. D., 2013: Projected effects of declining anthropogenic aerosols on the southern annular mode. Environ. Res. Lett., 8, 044028, https://doi.org/10.1088/1748-9326/8/4/044028.
Rotstayn, L. D., and Coauthors, 2014: Declining aerosols in CMIP5 projections: Effects on atmospheric temperature structure and midlatitude jets. J. Climate, 27, 6960–6977, https://doi.org/10.1175/JCLI-D-14-00258.1.
Rowell, D. P., 2019: An observational constraint on CMIP5 projections of the East African long rains and southern Indian Ocean warming. Geophys. Res. Lett., 46, 6050–6058, https://doi.org/10.1029/2019GL082847.
Rowell, D. P., and R. Chadwick, 2018: Causes of the uncertainty in projections of tropical terrestrial rainfall change: East Africa. J. Climate, 31, 5977–5995, https://doi.org/10.1175/JCLI-D-17-0830.1.
Santer, B. D., and Coauthors, 2013: Human and natural influences on the changing thermal structure of the atmosphere. Proc. Natl. Acad. Sci. USA, 110, 17 235–17 240, https://doi.org/10.1073/pnas.1305332110.
Shaw, T. A., and A. Voigt, 2016: Understanding the links between subtropical and extratropical circulation responses to climate change using aquaplanet model simulations. J. Climate, 29, 6637–6657, https://doi.org/10.1175/JCLI-D-16-0049.1.
Shaw, T. A., and Coauthors, 2016: Storm track processes and the opposing influences of climate change. Nat. Geosci., 9, 656–664, https://doi.org/10.1038/ngeo2783.
Sherwood, S. C., S. Bony, O. Boucher, C. Bretherton, P. M. Forster, J. M. Gregory, and B. Stevens, 2015: Adjustments in the forcing-feedback framework for understanding climate change. Bull. Amer. Meteor. Soc., 96, 217–228, https://doi.org/10.1175/BAMS-D-13-00167.1.
Simpson, I. R., and L. M. Polvani, 2016: Revisiting the relationship between jet position, forced response, and annular mode variability in the southern midlatitudes. Geophys. Res. Lett., 43, 2896–2903, https://doi.org/10.1002/2016GL067989.
Smith, C. J., and Coauthors, 2018: Understanding rapid adjustments to diverse forcing agents. Geophys. Res. Lett., 45, 12 023–12 031, https://doi.org/10.1029/2018GL079826.
Staten, P. W., J. J. Rutz, T. Reichler, and J. Lu, 2012: Breaking down the tropospheric circulation response by forcing. Climate Dyn., 39, 2361–2375, https://doi.org/10.1007/s00382-011-1267-y.
Steptoe, H., L. J. Wilcox, and E. J. Highwood, 2016: Is there a robust effect of anthropogenic aerosols on the southern annular mode? J. Geophys. Res. Atmos., 121, 10 029–10 042, https://doi.org/10.1002/2015JD024218.
Steptoe, H., S. E. O. Jones, and H. Fox, 2018: Correlations between extreme atmospheric hazards and global teleconnections: Implications for multihazard resilience. Rev. Geophys., 56, 50–78, https://doi.org/10.1002/2017RG000567.
Stjern, C. W., and Coauthors, 2017: Rapid adjustments cause weak surface temperature response to increased black carbon concentrations. J. Geophys. Res. Atmos., 122, 11 462–11 481, https://doi.org/10.1002/2017JD027326.
Tamarin-Brodsky, T., and Y. Kaspi, 2017: Enhanced poleward propagation of storms under climate change. Nat. Geosci., 10, 908–913, https://doi.org/10.1038/s41561-017-0001-8.
Tang, T., and Coauthors, 2019: Comparison of effective radiative forcing calculations using multiple methods, drivers, and models. J. Geophys. Res. Atmos., 124, 4382–4394, https://doi.org/10.1029/2018JD030188.
Thompson, D. W. J., and S. Solomon, 2002: Interpretation of recent Southern Hemisphere climate change. Science, 296, 895–899, https://doi.org/10.1126/science.1069270.
Voigt, A., and T. A. Shaw, 2015: Circulation response to warming shaped by radiative changes of clouds and water vapour. Nat. Geosci., 8, 102–106, https://doi.org/10.1038/ngeo2345.
Voigt, A., and T. A. Shaw, 2016: Impact of regional atmospheric cloud radiative changes on shifts of the extratropical jet stream in response to global warming. J. Climate, 29, 8399–8421, https://doi.org/10.1175/JCLI-D-16-0140.1.
Voigt, A., N. Albern, and G. Papavasileiou, 2019: The atmospheric pathway of the cloud-radiative impact on the circulation response to global warming: Important and uncertain. J. Climate, 32, 3051–3067, https://doi.org/10.1175/JCLI-D-18-0810.1.
Yin, J. H., 2005: A consistent poleward shift of the storm tracks in simulations of 21st century climate. Geophys. Res. Lett., 32, L18701, https://doi.org/10.1029/2005GL023684.
Zhao, X., R. J. Allen, T. Wood, and A. C. Maycock, 2020: Tropical belt width proportionately more sensitive to aerosols than greenhouse gases. Geophys. Res. Lett., 47, e2019GL086425, https://doi.org/10.1029/2019GL086425.
