1. Introduction
In the context of global warming, the impacts of polar sea ice changes, especially Arctic sea ice change, on weather and climate have received increasing attention in recent years (e.g., Liu et al. 2012; Cohen et al. 2014; Overland et al. 2019; Song et al. 2021). In contrast to the continued decline of Arctic sea ice in the past several decades, the Antarctic sea ice extent shows a significant increasing trend from 1979 to 2014, followed by an abnormal reduction in recent years (Turner et al. 2015, 2017; Parkinson 2019; Eayrs et al. 2021). A handful of modeling studies have investigated the impacts of Antarctic sea ice changes. Based on the observed Antarctic sea ice change, Raphael et al. (2011) showed that increased Antarctic sea ice might lead to a contraction of the Southern Hemisphere polar cell in summer associated with shifts in the Ferrel cell, and vice versa. Associated with increased Antarctic sea ice, Kidston et al. (2011) found that there is a significant poleward shift of the midlatitude jet in winter. Smith et al. (2017) also showed that increased Antarctic sea ice might drive a poleward shift of the midlatitude jet during the cold season. According to future projections of phases 5 and 6 of the Coupled Model Intercomparison Project (CMIP5/CMIP6), most climate models project a large reduction of Antarctic sea ice with increased greenhouse gases, even though the observed increasing trend during the period 1979–2014 is not well captured (Turner et al. 2013; Shu et al. 2015; Roach et al. 2020). For CMIP6, comparing with the multimodel mean during the period 1979–2014 from the historical simulation, Antarctic sea ice area will decrease by 90% in February and 50% in September under the highest emission scenario of the shared socioeconomic pathway (SSP)5-8.5 by the end of the twenty-first century (Roach et al. 2020). Based on the future Antarctic sea ice change, Bader et al. (2013) and England et al. (2018) suggested that there might be an equatorward shift of the winter atmospheric jet and storm tracks in the Southern Hemisphere associated with Antarctic sea ice loss under a high-emission scenario. Ayres and Screen (2019) indicated that extreme Antarctic sea ice loss under quadrupled CO2 might lead the eddy-driven jet in the Southern Hemisphere to more likely weaken than shift equatorward. More recently, Ayres et al. (2022) revealed that there is a weakening and equatorward shift of eddy-driven jet in response to Antarctic sea ice loss in both atmospheric and coupled model simulations, but the responses have greater magnitude in the coupled model simulation. Meanwhile, they also showed an increase in the mean temperature and precipitation in the Southern Hemisphere, but the responses are larger in the coupled model simulation. The above studies have suggested that Antarctic sea ice changes have impacts on the midlatitude atmospheric circulation and mean state of temperature and precipitation in the Southern Hemisphere. However, little is known about whether Antarctic sea ice loss can change temperature and precipitation extremes over the southern continents, which is the first focus of this study.
Recent research has revealed that ocean–atmosphere coupling might influence the magnitude of the mean atmospheric response in the Southern Hemisphere to projected loss of Antarctic sea ice and facilitate the response getting into lower latitudes. England et al. (2020b) indicated that ocean–atmosphere coupling might play an important role in facilitating the response to Antarctic sea ice loss getting deep into the tropics. Comparing the coupled and atmospheric model simulations, Ayres et al. (2022) further showed that the mean temperature and precipitation induced by Antarctic sea ice loss are largely confined to Antarctica in the atmospheric model simulation but are enhanced along with far-field responses in the coupled simulation. However, little is known about whether ocean–atmosphere coupling plays an important role in temperature and precipitation extremes in response to a large reduction of Antarctic sea ice, which is the second focus of this study.
As reported in the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR6) (Seneviratne et al. 2021), it is virtually certain that there have been increased warm extremes and decreased cold extremes on the global scale since 1950. It is likely that there have been increased heavy precipitation events since 1950. These changes would intensify with elevating global warming. Based on model projections involved in CMIP5/CMIP6, many studies have reached a similar conclusion that we will see increases in the occurrence of warm extremes, decreases in the occurrence of cold extremes, and increases in the occurrence of wet extremes (Fischer and Knutti 2015; Borodina et al. 2017; Sillmann et al. 2019; Di Luca et al. 2020; Li et al. 2021). Thus, a logical question is whether a large reduction of Antarctic sea ice makes a noticeable contribution to changes in temperature and precipitation extremes due to global warming.
This paper, for the first time, quantifies changes in temperature and precipitation extremes over the southern extratropical continents arising from Antarctic sea ice loss. Specifically, we want to answer the following three scientific questions:
-
How do temperature and precipitation extremes over the southern extratropical continents change in response solely to the projected different amount of reduction of Antarctic sea ice in fully coupled model experiments?
-
Do ocean dynamics and their interactions with the atmosphere matter for changes in temperature and precipitation extremes over the southern extratropical continents?
-
To what extent can changes in temperature and precipitation extremes over the southern extratropical continents associated with global warming be explained by Antarctic sea ice loss?
2. Methods
a. Model and experiment design
1) Coupled model experiments
The Community Earth System Model, version 1.2 (CESM1.2), which is a fully coupled model developed by the National Center for Atmospheric Research (NCAR), is employed to conduct coupled model experiments. It includes five model components: the Community Atmosphere Model, version 5 (CAM5), Community Land Model, version 4.0 (CLM4.0), Community Ice Code (CICE4), Parallel Ocean Program (POP2), and River Transport Model (RTM), which are coupled by the coupler (CPL7). More detailed information of CESM1.2 is documented in Hurrell et al. (2013). In this study, we run CESM1.2 at a horizontal resolution of approximately 1.9° latitude × 2.5° longitude for CAM5 and CLM4.0, 0.5° latitude × 1° longitude for POP2 and CICE4, and 0.5° latitude × 0.5° longitude for RTM. There are 30 vertical layers in the atmosphere and 60 vertical layers in the ocean.
Sun et al. (2020) and England et al. (2022) discussed different ways to prescribe Arctic sea ice loss in coupled climate models. In two recent coupled model studies that examine the effects of Antarctic sea ice loss on the mean climate by considering interactions from ocean dynamics with the atmosphere, the annual cycle of Antarctic sea ice cover is changed indirectly through modulating the Antarctic energy balance. One method is to specify an artificially seasonally varying downward longwave radiation and the other method is to reduce sea ice albedo. The artificial flux indirect method assumes that change in sea ice cover has a linear relationship with change in downwelling longwave radiation, but such relationship does not hold in summer. The albedo indirect method is not effective in adjusting year-round sea ice change because of little solar radiation during the polar night. Thus, it is hard to separate and identify the role of Antarctic sea ice loss precisely in the coupled climate model through these methods.
In this study, coupled model experiments based on CESM are performed by directly changing sea ice cover to directly assess whether different amounts of Antarctic sea ice loss have significant impacts on extreme temperature and precipitation over the southern extratropical continents. Specifically, we run three experiments in which Antarctic sea ice concentrations are prescribed with different repeating annual cycles. The radiative forcings (i.e., greenhouse gases and ozone) are fixed at the level of the year 2000 in all three experiments. Figure 1 shows the fixed seasonal cycle of Antarctic sea ice extent as well as the spatial distribution of seasonal Antarctic sea ice concentration for these experiments. In the first/control experiment (ASIlate20), the prescribed Antarctic sea ice cover is generated from the average of the historical simulations of CESM2 (included in CMIP6) during the period 1980–99. In ASIlate20, the prescribed Antarctic sea ice cover is similar to the observed Antarctic sea ice cover averaged between 1981 and 2010. In the second experiment (ASIlate21), the prescribed Antarctic sea ice cover is generated from the average of the SSP5-8.5 simulations of CESM2 (included in CMIP6) during the period 2080–99. SSP5-8.5 represents a high-emission scenario (Riahi et al. 2017). Compared with the control experiment, the second experiment is used to examine the impacts of the projected decreased Antarctic sea ice on temperature and precipitation extremes during the late twenty-first century. In the third experiment (ASI4CO2), we utilize a more extreme scenario, 1pctCO2-4xext of CESM2 (included in CMIP6), which is branched from a 1% yr−1 increase in atmospheric CO2 concentration run at year 140 and then run with atmospheric CO2 concentration fixed at 4 times the preindustrial level. In ASI4CO2, the prescribed Antarctic sea ice cover is generated from the average of the model year between 331 and 350. As shown in Fig. 1, the Antarctic is almost ice free for all months under atmospheric CO2 that is 4 times the preindustrial level. Compared with the control experiment, the third experiment is used to understand the effects of nearly ice-free Antarctic on temperature and precipitation extremes.
The prescribed seasonal Antarctic sea ice concentrations (upper panels) and seasonal cycle sea ice extent (lower panel) fixed in the three coupled model experiments. (a)–(d) Sea ice concentration averaged for September–November (SON), December–February (DJF), March–May (MAM), and June–August (JJA) in ASIlate20. (e)–(h) Seasonal averaged sea ice concentration in ASIlate21. (i)–(l) Seasonal averaged sea ice concentration in ASI4CO2. (m) Sea ice extent. The thick blue line in (a)–(l) represents sea ice edge defined as the contour of 15% ice concentration.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
In our coupled model experiments, we only fix Antarctic sea ice concentration at the end of the execution of the sea ice model component in CESM1.2. In this way, other sea ice variables (i.e., sea ice thickness) in the sea ice model component of CESM1.2 are allowed to evolve dynamically based on the forcings received from the ocean and atmosphere from the coupler. Thus, the oceanic and atmospheric model components of CESM1.2 can see the fixed Antarctic sea ice concentration but other sea ice variables are varied. Such a model setting allows 1) ocean–atmosphere interactions and ocean dynamics outside the region with the fixed Antarctic sea ice cover, 2) ocean dynamics below the fixed Antarctic sea ice cover, and 3) some feedbacks back to Antarctic sea ice. Such a model setting assumes that the simulated Antarctic sea ice cover by the sea ice model would gradually approach the fixed sea ice cover after a sufficient long integration. The simulation shows that ocean dynamics (i.e., AMOC) experience a large adjustment to the fixed Antarctic sea ice during the first 150-yr integration. After that, the AMOC reaches an approximate equilibrium. Using ASIlate21 as an example, as shown in Fig. S1 in the online supplemental material, the simulated sea ice cover by the sea ice model component of CESM1.2 agrees very well with the fixed ice cover after ocean dynamics reach an approximate equilibrium. We also plot the spatial distribution of Antarctic sea ice thickness simulated by the sea ice model component of CESM1.2 associated with the forcings received from the ocean and atmosphere. It appears that the simulated sea ice thickness also reaches an equilibrium after the adjustment of ocean dynamics (Fig. S2).
All simulations are run for 500 years. In this study, we select 301–500 years when the model is in an equilibrium state for analysis. Since the fixed Antarctic sea ice cover is repeated annually, the 301–500 years can be considered as 200 ensemble members. To analyze changes in temperature and precipitation extremes in response to Antarctic sea ice loss, we output daily temperature and precipitation for the three coupled model experiments. As discussed above, the only difference in the three coupled model experiments is the prescribed climatological Antarctic sea ice cover. Thus, the differences between the simulations of ASIlate21 and ASI4CO2, and the simulation of ASIlate20, represent the changes solely induced by the different amount of the reduction of Antarctic sea ice, rather than other factors (i.e., radiative forcings).
2) Atmospheric model experiments
The above three experiments are conducted using CEMS1.2 in the coupled configuration. That means ocean dynamics and their interactions with the atmosphere are allowed in response to Antarctic sea ice loss. To further explore the role of ocean dynamics and their interactions with the atmosphere in response to Antarctic sea ice loss in temperatures and precipitation extremes in the Southern Hemisphere, we conduct an additional two experiments using the atmospheric model component of CESM1.2—CAM5. Here, we run the CAM5 experiments with the same spatial resolution as the coupled model experiments. For atmospheric model simulations, the control experiment (CAM5_ASIlate20) uses the fixed Antarctic sea ice cover from the average of the historical simulations of CESM2 during the period 1980–99 (Figs. 1a–d), the observed sea surface temperature from the Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST1) (Rayner et al. 2003), and Arctic sea ice cover from the National Snow and Ice Data Center (Comiso 2017) averaged from 1981 to 2010 as the boundary condition. For the sensitivity experiment (CAM5_ASIlate21), first we fix Antarctic sea ice cover using the average sea ice concentration from the average of the SSP5-8.5 simulations of CESM2 during the period 2080–99 (Figs. 1e–h). Second, sea surface temperature is set to those in CAM5_ASIlate20 (the average of HadISST1 during the period 1981–2010) to isolate the impact of the Antarctic sea ice changes. Third, for grid cells where Antarctic sea ice concentration is inconsistent with that of the CAM5_ASIlate20, sea surface temperature is changed to the value obtained from the corresponding coupled model experiment (ASIlate21). Thus, only changes in sea ice cover and sea surface temperature directly associated with Antarctic sea ice changes are taken into account. Following similar approaches in Peings and Magnusdottir (2014) and England et al. (2018), sea ice thickness is set to 1 m in the Antarctic. Consistent with the coupled model experiments, radiative forcings of the CAM5 experiments are fixed at the level of the year 2000. Both CAM5 experiments are run for 111 years, and the last 110 years are for analysis.
b. Extreme indices
In this study, we select 10 indices as the measures of climate extremes from the core indices recommended by the World Meteorological Organization (WMO) Expert Team on Climate Change Detection and Indices (ETCCDI) (Zhang et al. 2011), which have been widely used in many studies (Sillmann et al. 2013; Screen et al. 2015; Song et al. 2021). Specifically, we generate six temperature and four precipitation extreme indices using daily model outputs of maximum (Tmax) and minimum (Tmin) surface temperature and total precipitation (Ptot), including cold days, cold nights, cold-spell duration, warm days, warm nights, heatwave duration, wet days, longest wet spell, dry days, and longest dry spell. Detailed definitions are listed in Table 1. Here, the 10th and 90th percentile threshold is calculated based on the daily outputs of the ASIlate20 experiment for coupled model experiments and CAM5_ASIlate20 for atmospheric model experiments.
Definition of extreme temperature and precipitation indices. Winter is defined as June, July, and August, and summer is defined as December, January, and February.
Based on domains defined in the IPCC Special Report for Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX) (Seneviratne et al. 2012) and the difference of spatial distribution of mean temperature and precipitation extremes between ASIlate21/ASI4CO2 and ASIlate20, we divide continents in the Southern Hemisphere into nine regions, including west of tropical Africa (AF1), east of tropical Africa (AF2), south of Africa (AF3), north of Australia (AU1), south of Australia (AU2), Peru and west of Brazil (SA1), east of Brazil (SA2), south of South America (SA3), and Antarctica (ANT) (Fig. 2). The aforementioned indices are calculated at each land grid and then averaged for each region. Here, we only focus on the changes in these indices that exceed the 95% significance level.
Nine regions defined in this study and named as follows: 1) west of tropical Africa (AF1; 11.4°S–0°, 5°–25°E), 2) east of tropical Africa (AF2; 11.4°S–0°, 25°–52°E), 3) south of Africa (AF3; 35°–11.4°S, 5°–52°E), 4) north of Australia (AU1; 30°–10°S, 110°–155°E), 5) south of Australia (AU2; 50°–30°S, 110°E–180°), 6) Peru and west of Brazil (SA1; 20°S–0°, 90°–50°W), 7) east of Brazil (SA2; 20°S–0°, 50°–34°W), 8) south of South America (SA3; 56.7°–20°S, 90°–39.4°W), and 9) Antarctica (ANT; 90°–60°S, 0°–0°).
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
c. Dynamical adjustment based on constructed analogs
To explore whether atmospheric circulation change plays a role in changes in extremes caused by Antarctic sea ice loss, a dynamical adjustment method is used. The objective of the dynamical adjustment method is to extract the component of specific physical variable variability that is only due to atmospheric circulation change. As described in Deser et al. (2016a), the dynamical adjustment method is based on a circulation analog reconstruction. More recently, using this method, Terray (2021) quantified the contribution of atmospheric circulation to two extreme events that occurred in 2012. Chripko et al. (2021) showed that the temperature response over some regions to reduced Arctic sea ice can be partly explained by changes in atmospheric circulation. Following these applications, here we applied the dynamical adjustment method to daily model outputs from our Antarctic sea ice loss experiments to derive the contribution of atmospheric circulation changes in the extreme responses. Like Terray (2021) and Chripko et al. (2021), daily sea level pressure (SLP) was used to represent atmospheric circulation changes. Daily Tmax, Tmin, and Ptot are physical variables that we focused on in this paper. Use Tmax as an example. To estimate the dynamical contribution of Tmax over the continents in the extratropical Southern Hemisphere, we selected SLP domains, which enclose land regions defined in the paper. For any day (Di) for each year in the ASIlate21/ASIlate20 experiment, the closest daily SLP analogs (Am) in all years within a time window of ±A days centered on day Di were selected from the ASIlate21/ASIlate20 experiment based on the Teweles–Wobus score (Terray 2021). We then randomly selected An out of Am (SLP analogs) and calculated the target SLP for day Di using the linear regression. The regression coefficients are applied to corresponding Tmax of An to estimate Tmax for day Di associated with atmospheric circulation. We repeated the random subsampling procedure 50 times. Finally, we averaged 50 reconstructed daily Tmax to get a dynamical component of daily Tmax in ASIlate21/ASIlate20. Here, A = 5, Am = 300, An = 150. The difference in the dynamical component of daily Tmax between ASIlate21 and ASIlate20 represents Tmax changes due to SLP changes caused by Antarctic sea ice loss. The residual part mainly represents changes corresponding to thermodynamic processes (i.e., advection of anomalous air mass over the ocean by the climatological flow, local heat flux changes; Chripko et al. 2021).
d. CESM1 large ensemble
The CESM1 Large Ensemble Community Project (LENS) is also used to facilitate the analysis (including 40 members; Kay et al. 2015), given that CESM2 LENS does not provide the output under the SSP5-8.5 scenario. Specifically, the CESM1 LENS under the historical simulation and the projection under the RCP8.5 emission scenario is used. The sea ice concentration averaged during the period 1980–99 and during the period 2080–99 in CESM1 LENS is similar to our prescribed ice concentration derived from CESM2. Consistent with the time periods of ASIlate20 and ASIlate21, we calculate temperature and precipitation extremes from the historical simulation for the period 1980–99 (hereinafter referred to as LENSlate20) and from the projection for the period 2080–99 (hereinafter referred to as LENSlate21), respectively. For LENSlate20 and LENSlate21, the 10th, 90th percentile threshold is calculated based on LENSlate20. It provides us an approximate estimation of the possible contribution of Antarctic sea ice loss to changes in temperature and precipitation extremes due to greenhouse warming, by comparing the changes of temperature and precipitation extremes between our coupled model experiments as described above and the CESM1 large ensemble experiments.
3. Results
a. Changes in mean state and variance
Associated with decreased Antarctic sea ice by the end of the twenty-first century, ASIlate21 shows significant annual mean warming for all continents in the extratropical Southern Hemisphere (Fig. 3a). The warming tends to be stronger at higher latitudes than at lower and midlatitudes, and the largest warming anomaly is found over Antarctica. The temperature increases by 1°–2°C over Africa, Australia, and South America, and by ∼4°C over Antarctica. Unlike the uniform mean warming, small reduced temperature variance is projected over midlatitude regions (AU2 and SA3, Fig. 3b), and small increased variance occurs over most tropical and subtropical regions. In particular, Antarctica experiences a large variance increase. The responses to nearly ice-free Antarctic in ASI4CO2 are consistent with ASIlate21, but the magnitude is larger by more than 30% relative to those of ASIlate21 (Figs. S3a,b).
Regional changes of annual-mean temperature and precipitation and their daily variance in response to ASIlat21 and CAM5_ASIlate21. (a) Annual-mean temperature (°C), (b) daily temperature variance (°C2), (c) annual-mean total precipitation (mm day−1), and (d) daily precipitation variance (mm2 day−2); (a)–(d) are responses to ASIlat21. (e)–(h) As in (a)–(d), but responses are to CAM5_ASIlate21. Color-filled bars are statistically significant at the 95% confidence level, whereas blank no-fill bars are not statistically significant.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
For precipitation, ASIlate21 produces a significant increase over most of the southern continents, except for no significant change over AF3 and AU2 (Fig. 3c). The largest increase in precipitation is found over SA2 (∼0.4 mm day−1). The increased precipitation is primarily in the form of rainfall at lower and midlatitudes (AF1, AF2, Australia, South America), except that there is a slight decrease in snowfall in SA3 in part due to the mean warming. By contrast, Antarctica sees increased precipitation mainly in the form of snow. Accompanying increased mean precipitation, ASIlate21 projects an increase in precipitation variance over most regions, except for no significant change over AF3 and AU2. Consistent with ASIlate21, ASI4CO2 shows increased rainfall with a larger magnitude over the continents in lower and midlatitudes, including AU2, and snowfall over Antarctica (Fig. S3c). ASI4CO2 also leads to an enhanced magnitude of increased precipitation variance (Fig. S3d).
As shown in Fig. 3e, in the absence of ocean dynamics and their interactions with the atmosphere, the temperature responses to CAM5_ASIlate21 are much weaker. Moreover, most regions in lower and midlatitudes have slight cooling that is not statistically significant. Significant warming is only found over SA2 and Antarctica due to the local response to sea ice loss (Fig. 3e). Meanwhile, the temperature variance has no significant change for all regions (Fig. 3f). For precipitation, the responses in the CAM5_ASIlate21 are also much weaker and insignificant compared with those in the coupled model simulation, and some regions have an opposite rainfall change. Only over Antarctica, there is a small but significant increased snowfall. Consistent with this small increased snowfall, there is a small increased precipitation variance over Antarctica (Fig. 3h), but it is much weaker than that in the coupled model simulation. Consistent with the result of Ayres et al. (2022), in the absence of air–sea coupling, large changes in the mean temperature and precipitation are mainly limited to southern high latitudes.
b. Changes in climate extremes
1) Temperature
As shown in Figs. 4a and 4b, ASIlate21 results in significant increased warm days and nights over all regions in summer. The increase of warm days is much larger over high-latitude regions than that over lower- and midlatitude regions, that is, Antarctica sees more than 20 warm days. By contrast, most regions show a much larger increase in warm nights (>20 days), except Australia (∼10 days). Africa and SA1 have more than 40 warm nights. Thus, the changes in warm nights are more robust than those in warm days. Coinciding with increased warm days and nights, projected decreased Antarctic sea ice also affects the heatwave duration. As shown in Fig. 4c, there is a small but significant increase of heatwave duration over lower- and midlatitude regions. Antarctica has the largest increase of heatwave duration (∼18 days), which is ∼4–5 times larger than other regions. As for cold extremes in winter, in ASIlate21, almost all regions experience significant decreased cold days and nights, except for no significant change of cold days over SA2 (Figs. 4d,e). The largest decrease in cold days is found over Antarctica (followed by AU2), while the decrease of cold nights is relatively uniform over all regions. The magnitude of the change in cold days and nights is smaller than those of warm days and nights. This is partly due to a potential limitation of the cold extremes that we used, that is, cold days cannot be less than 0 days. In line with the decrease in cold days and nights, there is a small but significant shift toward a shortened cold-spell duration over all regions. Associated with a further reduction of Antarctic sea ice in ASI4CO2, the responses of warm and cold extremes are generally consistent with those in ASIlate21, but the magnitude is enhanced for warm extremes over all regions, that is, nearly doubled for warm days and nights and heatwave duration, and relatively enhanced for cold extremes for most regions (Fig. S4).
Regional changes of temperature extremes in response to ASIlate21 and CAM5_ASIlate21. (a) Warm days in summer, (b) warm nights in summer, (c) heatwave duration in summer (d) cold days in winter, (e) cold nights in winter, and (f) cold-spell duration in winter; (a)–(f) are responses to ASIlate21. (g)–(l) As in (a)–(f), but responses are to CAM5_ASIlate21. Color-filled bars are statistically significant at the 95% confidence level, whereas blank no-fill bars are not statistically significant.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
As mentioned previously, the results of England et al. (2018) and Ayres et al. (2022) indicated that in the absence of interactions of ocean dynamics with the atmosphere, large changes in the mean temperature and precipitation are mainly confined to southern high latitudes. Next, we compare the changes of climate extremes in the atmosphere-only model in response to Antarctic sea ice loss with those in the coupled model. As shown in Figs. 4g–l, without ocean dynamics and their interactions with the atmosphere, the CAM5_ASIlate21 produces much smaller changes in warm and cold extremes than those in the coupled CESM simulation (ASIlate21). In the CAM5 experiment, there is no significant change of warm (cold) extremes for all (almost all) regions (Figs. 4g–l), and the sign of the changes is not uniform across the regions. Only Antarctica shows significant decreased cold days and nights and cold-spell duration (Figs. 4j–l). Figure S5 shows the spatial distribution of changes in the mean temperature and temperature extremes over Antarctica. Significant mean warming found over Antarctica in the atmospheric model simulation is largely contributed by the mean warming over west Antarctica. There is almost no significant change in warm extremes (including warm days, warm nights, and heat wave duration) found over Antarctica, but significant decreased cold extremes (including cold days, cold nights, and cold-spell duration) are found over west Antarctica. Thus, the significant mean warming over west Antarctica is contributed more by decreased cold extremes than by changes in warm extremes. This illustrates that potential feedbacks from interactions of ocean dynamics with the atmosphere induced by the reduction of Antarctic sea ice play an important role in the changes in temperature extremes, which leads to robust and significant changes in warm and cold extremes over the extratropical Southern Hemisphere continents.
2) Precipitation
Associated with reduced Antarctic sea ice in ASIlate21, in summer, most regions have increased wet days, but there is no significant change over AF2, AF3, and SA1 (Fig. 5a). The largest increase of wet days occurs over SA2 (∼3 days). The longest wet-spell duration has significantly increase over AF3, Australia, SA2, SA3, and Antarctica (Fig. 5b). By contrast, it decreases over AF1 and AF2. As for dry days in summer, ASIlate21 results in a significant decrease over AU1, SA3, and Antarctica and shows no significant change over the rest of the regions (Fig. 5c). The magnitude of the changes of dry days (less than 0.5 days) is much smaller than that of wet days. The longest dry-spell duration is projected to lengthen over AF2, SA1, and Antarctica (Fig. 5d). Over Antarctica, although there are decreased dry days, the longest dry-spell duration is projected to increase. The longest dry-spell duration becomes shorter over AU1 and SA3 (consistent with the change of dry days), as well as AU2 and SA2.
Regional changes of precipitation extremes in response to ASIlate21 and CAM5_ASIlate21. (a) Wet days in summer, (b) longest wet spell in summer, (c) dry days in summer, (d) longest dry spell in summer. (e)–(h) As in (a)–(d), but for winter; (a)–(h) are responses to ASIlate21. (i)–(p) As in (a)–(h), but responses are to CAM5_ASIlate21. Color-filled bars are statistically significant at the 95% confidence level, whereas blank no-fill bars are not statistically significant.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
In winter, ASIlate21 projects small but significant increases in wet days over AF2, South America, and Antarctica, and no significant change over the rest of the regions (Fig. 5e). The increase of wet days over Antarctica is larger than those of other regions. Consistent with the changes in wet days, an increased longest wet-spell duration is also found over South America and Antarctica (Fig. 5f), but the largest change is over SA1. For dry days, only three regions experience significant changes. AF2 and AF3 have more dry days, whereas SA1 has fewer dry days (Fig. 5g). The longest dry-spell duration over AF2, Australia, and SA2 becomes longer (Fig. 5h).
The further reduction of Antarctic sea ice in ASI4CO2 results in the changes of wet and dry extremes that are broadly similar to those in ASIlate21, but with larger changes over most regions, especially in winter (Fig. S6).
Without ocean dynamics and their interactions with the atmosphere, the atmosphere-only model simulation (CAM_ASIlate21) produces much smaller changes in wet and dry extremes than those in the coupled model simulation (ASIlate21). In the CAM5 simulation, there is no significant change of wet and dry extremes in summer for all regions (Figs. 5i–l), and the signs of the changes are not consistent with those found in the coupled model simulation. This is also the case for wet and dry extremes in winter, except Antarctica has significant increased wet days and longest wet spell (Figs. 5m,n). SA3 has a significant increase in the longest dry spell, while AU1 and ANT see opposite changes. This further illustrates the importance of air–sea coupling associated with Antarctic sea ice loss in the changes in precipitation extremes over the extratropical Southern Hemisphere continents.
c. Comparison with extreme changes due to greenhouse warming
The above analyses show that a large reduction of Antarctic sea ice does have significant impacts on temperature and precipitation extremes over the Southern Hemisphere continents. It has been recognized that greenhouse warming can contribute substantially to temperature and precipitation extremes. Here, we further explore the potential contribution of Antarctic sea ice loss to the changes of temperature and precipitation extremes resulting from increased greenhouse gases.
As shown in Fig. 6, under the RCP8.5 scenario, all continents of the Southern Hemisphere are projected to have a mean warming of 4°–5°C by the end of the twenty-first century (Fig. 6a, red bars). This warming is comparable to the multimodel mean global warming by the end of the twenty-first century (Li et al. 2021). Over Antarctica, Antarctic sea ice loss makes a large contribution to the warming, whereas over other regions, the contribution is small. For precipitation, LENSlate21 projects significantly increased precipitation over most regions, except for decreased precipitation over SA1 and no significant change over AF3 (Fig. 6b, red bars). These changes are consistent with spital precipitation changes revealed by Lin et al. (2016). LENSlate21 also projects significant decreased snowfall over SA3 and increased snowfall over Antarctica. Antarctic sea ice loss contributes a large portion of increased precipitation over Antarctica, AU1, SA2, and SA3, but a small portion over Africa.
Regional averaged changes of annual-mean temperature and precipitation in response to ASIlate21 (blue bars) and LENSlate21 (red bars). (a) Annual-mean temperature and (b) annual-mean total precipitation.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
In LENSlate21, the summer warm days are projected to have a large increase over all regions, especially Africa and north of South America (>55 days, Fig. 7a). Antarctic sea ice loss makes a small contribution to the increased warm days over most regions, except for Antarctica, where it makes a noticeable contribution to increased warm days. The summer warm nights are projected to increase over all regions similar to those of warm days, but with a larger magnitude. Antarctic sea ice loss may contribute to a small portion of the increased warm nights (Fig. 7b). Coinciding with the changes of warm days and nights, the heatwave duration is projected to be longer over all regions, especially over Africa, South America, and Antarctica (Fig. 7c). Similar to warm days, Antarctic sea ice loss makes a small contribution to the increased heatwave duration over most regions, except Antarctica. As for winter cold extremes, in LENSlate21, the cold days and nights show similar changes over all regions. Both are projected to decrease by ∼8–9 days (Figs. 7d,e). That means cold days and nights will rarely occur over the Southern Hemisphere at the end of the twenty-first century, since there are about 9 cold days and nights in winter in LENSlate20. The decreased cold-spell duration is found over all regions except AF2 (Fig. 7f). Antarctic sea ice loss may contribute to a part of the decreased cold spell over these regions.
Regional averaged changes of temperature extremes in response to ASIlate21 and LENSlate21. (a) Warm days in summer, (b) warm nights in summer, (c) heatwave duration in summer (d) cold days in winter, (e) cold nights in winter, and (f) cold-spell duration in winter. Blue bars are responses in ASIlate21. Red bars are responses in LENSlate21.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
Compared with quite uniform changes in temperature extremes, changes in precipitation extremes vary greatly with regions (Fig. 8). In LENSlate21, the summer wet days are projected to increase over Africa, Australia, SA2, SA3, and Antarctica, and the largest increase is over AF1 (∼9 days) (Fig. 8a, red bars). Antarctic sea ice loss may make a large contribution to the increase over AU1 and SA2 and a small contribution over AF1, AU2, and SA3. The changes of the summer longest wet spell vary with regions. LENlate21 projects an increase over Australia, SA3, and Antarctica and a decrease over Africa, SA1, and SA2 (especially SA1, decreased by ∼3.5 days). The sea ice loss tends to dominate the changes over Africa, Australia, SA3, and Antarctica (Fig. 8b). The dry days are projected to increase over AF3, AU2, and SA3, but they are projected to decrease over Antarctica (Fig. 8c). The sea ice loss may play a role over Antarctica. The changes of longest dry spell also vary with regions. The changes of summer longest dry spell are projected to increase over AF1, AF3, SA1, and Antarctica (especially Antarctica, increased by ∼3.5 days) and decrease over AF2 and Australia. Antarctic sea ice loss may dominate the decreased longest dry spell over Australia and make a large contribution to increased longest dry spell over SA1 and Antarctica (Fig. 8d).
Regional averaged changes of precipitation extremes in response to ASIlate21 and LENSlate21. (a) Wet days in summer, (b) longest wet spell in summer, (c) dry days in summer, and (d) longest dry spell in summer. (e)–(h) As in (a)–(d), but for winter. Blue bars are responses in ASIlate21. Red bars are responses in LENSlate21.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
In winter, LENSlate21 projects increased wet days over AF1, AF2, SA1, SA2, and Antarctica, with a small decrease over AF3 (Fig. 8e). Antarctic sea ice loss may make a large contribution to the increase over SA2 and Antarctica and may play a role over SA1. The longest wet spell is projected to increase over AF1, AF2, and Antarctica and decrease over AF3, AU2, and South America (Fig. 8f). The sea ice loss only plays a dominant role over Antarctica. In LENSlate21, dry days are projected to increase over all regions (Fig. 8g). The sea ice loss only makes a small contribution to increased dry days over AF2 and AF3. The longest dry spell is projected to increase over most regions but decrease over AF3 and SA2 (Fig. 8h). Antarctic sea ice loss may dominate the increased longest dry spell over AF2 and AU1.
d. Mean atmospheric circulation changes
As shown in Figs. S7a and S7b, ASIlate21 leads to significant warming from the surface to the upper troposphere over the entire Southern Hemisphere for both summer and winter. This is in line with the warming revealed by England et al. (2020b) and Ayres et al. (2022). This response is also analogous to the “mini global warming” signature caused by Arctic sea ice loss simulated by the Community Climate System Model, version 4 (Deser et al. 2015, 2016b). Associated with the reduction of Antarctic sea ice, the Southern Ocean receives more incoming solar radiation, increasing sea surface temperature and heat stored in the upper ocean. In turn, the ocean is able to transfer more heat and moisture (sensible and latent heat flux) to the atmosphere. This results in intensified surface warming between ∼60° and ∼80°S, especially in winter (surface warming can be ∼4°C), and the warm anomalies decay from the surface to the upper troposphere and from high latitudes to the subtropics (Figs. 7a,b). Both observations (Westra et al. 2013) and model simulations (Kharin et al. 2013; Li et al. 2021) suggested that the mean precipitation tends to be proportional to the mean surface temperature change according to the Clausius–Clapeyron relationship. This hemispheric warming caused by Antarctic sea ice loss tends to increase precipitation in the Southern Hemisphere and thus, the wet extremes as discussed above. Consistent with the decreased temperature gradient between high and midlatitudes, the zonally averaged zonal wind exhibits weaker westerlies centered ∼60°S in both summer and winter (Figs. 9a,b). This implies a weakening of the eddy-driven jet as well as the polar front. In winter, ASIlate21 also leads to weaker westerlies south of 30°S and stronger westerlies north of 30°S, especially in the mid–upper troposphere. This implies a northward shift of the subtropical jet stream, which influences storm tracks. Associated with further reduction of Antarctic sea ice, ASI4CO2 leads to responses similar to those of ASIlate21, but with much larger anomalies.
Changes in zonal-mean of zonal wind in response to ASIlate21 and CAM5_ASIlate21. Responses in ASIlate21 (a) in summer and (b) in winter. (c), (d) As in (a) and (b), but responses are in CAM5_ASIlate21. Dotted areas indicate statistically significant at 95% confidence level.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
Figures 10a and 10b show sea level pressure and near-surface wind changes due to Antarctic sea ice loss in ASIlate21. There is below-normal pressure over the subtropics and midlatitudes and above-normal pressure over high latitudes. This is similar to the negative phase of the Antarctic Oscillation (AAO), which is the dominant climate mode in the extratropical Southern Hemisphere (Gong and Wang 1999; Jones and Widmann 2003). This is consistent with a weakening of the westerly flow that lies in the midlatitudes centered ∼60°S (Figs. 9a,b). The changes of AAO have a significant impact on temperature and precipitation over the Southern Hemisphere continents. Previous studies have shown that there is increased temperature over Australia and large parts of Antarctica and increased precipitation over Australia, south of Africa, and south of South America, associated with the negative AAO phase (Gillett et al. 2006; Pezza et al. 2008; Yu et al. 2012; Zheng et al. 2015; Babian et al. 2016). These changes are consistent with the increases of temperature and precipitation extremes analyzed above. We know that the midlatitude weather variability is induced by Rossby waves associated with variations of storms and synoptic time scale systems. Here, we apply a 2.5–6-day bandpass filter to zonal and meridional winds and then calculate the eddy kinetic energy (EKE) (Khokhlov et al. 2004). As shown in Figs. 10c and 10d, ASIlate21 results in decreased EKE in midlatitudes, especially in winter. This manifests the decreased weather variability, which may result in decreased cold extremes in winter.
Changes in SLP and near-surface wind and EKE at 850 hPa due to Antarctic sea ice loss in ASIlate21. Difference of SLP and wind between ASIlate21 and ASIlate20 in (a) summer and (b) winter. (c),(d) As in (a) and (b), but responses are of EKE. Dotted areas indicate statistically significant at 95% confidence level.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
In the absence of ocean–atmosphere coupling, in summer, the CAM5 simulation only shows very weak warming in the lower troposphere at high latitudes and in the midtroposphere at midlatitudes and cooling in the upper troposphere at the tropics (Fig. S7c). This is accompanied by slightly weaker westerlies around 30°S (Fig. 9c). In winter, there is an intensified local and shallow warming between 50° and 80°S (Fig. S7d). Meanwhile, there is a small but significant cooling in the midtroposphere between 80° and 90°S. The changes in temperature gradient result in weaker westerlies around 55°S and stronger westerlies around 75°S (Fig. 9d). These responses are generally consistent with the results of Kidston et al. (2011) and Bader et al. (2013).
Compared with the hemispherical warming and circulation changes caused by the Antarctic sea ice loss in the coupled model simulation, the atmospheric circulation responses in the atmosphere-only simulation are much smaller and are mainly constrained to mid–high latitudes. Correspondingly, the changes in temperature and precipitation extremes in the atmosphere-only model mainly occur over high latitudes. Both the coupled and atmospheric model show large increased heat flux from the ocean to atmosphere in the regions where sea ice is reduced (30–40 W m−2, Fig. S8). However, the coupled model simulation also shows significant increased heat flux from the ocean to atmosphere in the subtropical and midlatitude oceans (15–20 W m−2), which is absent in the stand-alone atmospheric model simulation (Fig. S8). Figure 11 shows the changes in northward atmospheric heat transport due to Antarctic sea ice loss in the coupled and atmospheric model simulations. The large northward atmospheric heat transport anomaly in the atmosphere-only model is confined to high latitudes, which is consistent with the results given by England et al. (2020b). By contrast, in the coupled model, there is enhanced northward atmospheric heat transport in the entire Southern Hemisphere, which is consistent with the above surface heat flux changes. This would influence various climate feedbacks (i.e., water vapor feedback, Liu et al. 2018) and thus result in changes in temperature and precipitation extremes.
Changes in northward atmospheric heat transport due to Antarctic sea ice loss in ASIlate21 and CAM5_ASIlate21. [calculated following the method in the appendix of Kay et al. (2012)].
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
e. Contribution of dynamical components to extreme changes
To explore whether atmospheric circulation change plays a role in changes in extremes caused by Antarctic sea ice loss, we apply the dynamical adjustment method to the coupled model simulation. As shown in Fig. S9, the dynamical component that reflects the effect of atmospheric circulation change due to Antarctic sea ice loss appears to account for a small to moderate portion of the increased summer Tmax over AU2, large parts of South America, and Antarctica (Fig. S9a), and a small to moderate portion of the increased summer Tmin over all subregions except Africa (Fig. S9b). As shown in Fig. S9c, only a nearly half (small) portion of the increased winter Tmax over Antarctica (SA3) can be explained by dynamical change, whereas a small to moderate portion of the increased winter Tmin over all subregions, except AF3 and AU1, can be attributed to the dynamical component (Figs. S9d,e). For changes in precipitation induced by Antarctic sea ice loss, the dynamical component shows varying contributions for each subregion in summer, but it plays a dominant role over most subregions in winter (Figs. S9c,f).
As shown in Figs. 12a and 12b, a half (moderate) portion of the increased summer warm days over SA1 (Antarctica) can be explained by dynamical change, whereas a small to moderate portion of the increased summer warm nights over South America and Antarctica can be attributed to the dynamical component. For cold extremes, the dynamical component accounts for a small to moderate portion of the reduced winter cold days over Africa, AU1, SA3, and Antarctica (Fig. 12c), and a small to large portion of the reduced winter cold nights over nearly all subregions except AF3 (Fig. 12d). As shown in Fig. 13, for precipitation extremes, the dynamical component shows varying contributions for different regions, and for some regions, the sign of the changes is opposite.
Regional averaged total and dynamical component responses of temperature extremes to Antarctic sea ice loss in ASIlate21. (a) Warm days in summer, (b) warm nights in summer, (c) cold days in winter, and (d) cold nights in winter. Blue bars are total responses. Light blue bars are dynamical component responses.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
Regional averaged total and dynamical component responses of precipitation extremes to Antarctic sea ice loss in ASIlate21. (a) Wet days in summer, (b) dry days in summer. (c),(d) As in (a) and (b), but for winter. Blue bars are total responses. Light blue bars are dynamical component responses.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
4. Discussion and conclusions
In this study, we investigate the impacts of projected future Antarctic sea ice loss on temperature and precipitation extremes over the southern continents. To isolate the impacts of Antarctic sea ice change, coupled model experiments are forced with different amounts of Antarctic sea ice cover but allow ocean dynamics and their interactions with the atmosphere induced by Antarctic sea ice change. We further conduct corresponding atmosphere-only experiments to determine whether ocean–atmosphere coupling plays an important role in changes in temperature and precipitation extremes in response to a large reduction of Antarctic sea ice. Here, we focus on the responses of temperature and precipitation extremes in summer and winter, but we also include the responses in spring and autumn in the supplemental material (Figs. S10 and S11).
For the mean state, our coupled model experiments show that Antarctic sea ice loss results in an increase in the mean temperature and precipitation over the southern extratropical continents. This is broadly consistent with recent modeling studies using different coupled models and sea ice constrain approaches. Using the ghost flux approach to constrain Antarctic sea ice in the CESM1 Whole Atmosphere Community Climate Model (WACCM4), England et al. (2020a,b) showed Antarctic sea ice loss leads to increased annual mean temperature over almost all land areas and precipitation over tropical and subtropical land areas in the Southern Hemisphere. Using the albedo reduction approach in the HadGEM3-GC3.1-LL model, Ayres et al. (2022) showed that a large reduction of Antarctic sea ice results in warming over the entire southern continents and precipitation over parts of the southern continents. We also notice that temperature variance is reduced over southern Australia and South America but is enhanced relatively over most tropical and subtropical regions and greatly over Antarctica. Precipitation variance is enhanced over most southern continents.
For climate extremes, we find that Antarctic sea ice loss that is projected by the end of the twenty-first century does have significant impacts on the frequency and duration of temperature and precipitation extremes over the southern continents. Given that several climate extreme indices and subregions are discussed in this study, for a clear take-home message, we summarize for which extreme and over which region there is a significant change induced by Antarctic sea ice loss in Fig. 14. In general, the reduction of Antarctic sea ice in ASIlate21 leads to a significant increase in the occurrence and duration of summer warm extremes over all regions and a decrease in the occurrence and duration of winter cold extremes over almost all regions. The occurrence and duration of wet extremes are projected to significantly increase over Australia, the east and south parts of South America, and Antarctica in summer, and over South America and Antarctica in winter. However, the changes of dry extremes vary with regions (sign and magnitude). Using a dynamical adjustment method based on the reconstruction of atmospheric circulation analogs, our analysis suggests that the temperature extremes over some regions may be partly explained by changes in the atmospheric circulation, whereas the precipitation extremes over north of South America may be largely attributed to the dynamic change. Compared with ASIlate21, further loss of Antarctic sea ice under a quadrupling of CO2 (ASI4CO2) leads to similar changes in temperature and precipitation extremes but with much larger anomalies. The results here are based on the CESM model. Coordinated experiments, including those that utilize different coupled climate models and different model configurations, are needed to further quantify changes in climate extremes over the southern extratropical continents in response to Antarctic sea ice loss. Reponses of extreme temperature and precipitation might be more related to the regions having the largest reduction in sea ice, (e.g., compared with ASIlate20, the largest reduction in sea ice in ASIlate21 is located in the east Atlantic and the eastern Pacific sectors, Fig. 1). Compared with ASIlate21, the ASI4CO2 experiment has a very different sea ice loss distribution (almost ice free for all months). However, ASI4CO2 leads to similar changes in temperature and precipitation extremes but with much larger anomalies. This in part suggests that the responses of extreme temperature and precipitation might be more related to changes in the entire Antarctic sea ice cover. The sensitivity to the regionality of sea ice loss will be explored in future research.
Temperature and precipitation extremes in response to Antarctic sea ice loss at end of twenty-first century. Red boxes indicate significant increase; blue boxes indicate significant decrease; white boxes indicate no significant response.
Citation: Journal of Climate 36, 14; 10.1175/JCLI-D-22-0577.1
Some studies have examined possible impacts of decreased Antarctic sea ice cover on atmospheric circulation in the Southern Hemisphere by forcing atmosphere-only models with historical and projected sea ice changes. Most of them found a weakening of midlatitude jet (Menéndez et al. 1999; Bader et al. 2013; England et al. 2018), although Kidston et al. (2011) suggested no significant change. Based on the CMIP5 models, Ayres and Screen (2019) demonstrated that atmospheric responses to loss of Antarctic sea ice are largely confined to the southern high-latitude and lower atmosphere. The results of our uncoupled model experiments are generally consistent with these studies, and further showed that significant changes in temperature and precipitation extremes are only found over Antarctica.
Recently, the simulation of England et al. (2020b) indicated that the ocean–atmosphere coupling might play an important role in facilitating the response to Antarctic sea ice loss getting deep into the tropics. Comparing coupled and uncoupled simulations based on the Hadley Center model, Ayres et al. (2022) suggested that ocean–atmosphere coupling associated with the loss of Antarctic sea ice can result in the mean atmospheric responses with greater magnitude and generate far-field responses. Comparison of our coupled and atmosphere-only model experiments further showed that with ocean dynamics and their interactions with the atmosphere, Antarctic sea ice loss produces much larger and significant changes in the mean temperature (including Tmax and Tmin) and precipitation over the southern extratropical continents.
More importantly, in contrast to minor and insignificant changes in our atmosphere-only model experiments, we find that with ocean dynamics and their interactions with the atmosphere in our coupled model experiments, Antarctic sea ice loss produces a significant increase in the frequency and duration of warm extremes over the southern continents, a decrease in cold extremes over most regions, an increase in wet extremes over South America and Antarctica, and changes in dry extremes (but vary by region). This highlights that potential feedbacks from ocean–atmosphere coupling induced by Antarctic sea ice loss play a critical role in driving significant changes in climate extremes over the southern continents.
Further comparing with global warming experiments that include the impacts of greenhouse gas forcing and sea ice loss (both Antarctic and Arctic), we find that Antarctic sea ice loss may make a significant contribution to temperature extremes for most regions but a varying contribution (sign and magnitude) to precipitation across the regions. However, it should be noted that the 20-yr period from the historical and projection simulations are transient simulations, although the changes in extremes between them can be considered as an approximate steady-state response to greenhouse warming based on the large ensemble. Thus, the identified contribution of Antarctic sea ice loss to changes in extremes due to greenhouse warming might be relatively overestimated. Future research is needed to better quantify this.
Acknowledgments.
This research is supported by the National Natural Science Foundation of China (41830536 and 41941009) and the National Key R&D Program of China (2018YFA0605901). We thank the National Center for Atmospheric Research (NCAR) for making data of the CESM Large Ensemble Community Project publicly available. We also thank the three anonymous reviewers for their constructive comments.
Data availability statement.
Data of the CESM Large Ensemble Community Project can be downloaded at https://www.cesm.ucar.edu/projects/community-projects/LENS/. The output of model simulations conducted in this paper can be accessed online at https://doi.org/10.5281/zenodo.7635716.
REFERENCES
Ayres, H. C., and J. A. Screen, 2019: Multimodel analysis of the atmospheric response to Antarctic sea ice loss at quadrupled CO2. Geophys. Res. Lett., 46, 9861–9869, https://doi.org/10.1029/2019GL083653.
Ayres, H. C., J. A. Screen, E. W. Blockley, and T. J. Bracegirdle, 2022: The coupled atmosphere–ocean response to Antarctic sea ice loss. J. Climate, 35, 4665–4685, https://doi.org/10.1175/JCLI-D-21-0918.1.
Babian, S., H. W. Rust, J. Grieger, K. Prömmel, and U. Cubasch, 2016: Representation of the Antarctic Oscillation and related precipitation patterns in the MPI Earth System Model. Meteor. Z., 25, 767–774, https://doi.org/10.1127/metz/2016/0661.
Bader, J., M. Flügge, N. G. Kvamstø, M. D. S. Mesquita, and A. Voigt, 2013: Atmospheric winter response to a projected future Antarctic sea-ice reduction: A dynamical analysis. Climate Dyn., 40, 2707–2718, https://doi.org/10.1007/s00382-012-1507-9.
Borodina, A., E. M. Fischer, and R. Knutti, 2017: Potential to constrain projections of hot temperature extremes. J. Climate, 30, 9949–9964, https://doi.org/10.1175/JCLI-D-16-0848.1.
Chripko, S., R. Msadek, E. Sanchez-Gomez, L. Terray, L. Bessières, and M.-P. Moine, 2021: Impact of reduced Arctic sea ice on Northern Hemisphere climate and weather in autumn and winter. J. Climate, 34, 5847–5867, https://doi.org/10.1175/JCLI-D-20-0515.1.
Cohen, J., and Coauthors, 2014: Recent Arctic amplification and extreme mid-latitude weather. Nat. Geosci., 7, 627–637, https://doi.org/10.1038/ngeo2234.
Comiso, J. C., 2017: Bootstrap sea ice concentrations from Nimbus-7 SMMR and DMSP SSM/I-SSMIS, version 3. NASA National Snow and Ice Data Center, accessed 17 August 2019, https://doi.org/10.5067/7Q8HCCWS4I0R.
Deser, C., R. A. Tomas, and L. Sun, 2015: The role of ocean–atmosphere coupling in the zonal-mean atmospheric response to Arctic sea ice loss. J. Climate, 28, 2168–2186, https://doi.org/10.1175/JCLI-D-14-00325.1.
Deser, C., L. Terray, and A. S. Phillips, 2016a: Forced and internal components of winter air temperature trends over North America during the past 50 years: Mechanisms and implications. J. Climate, 29, 2237–2258, https://doi.org/10.1175/JCLI-D-15-0304.1.
Deser, C., L. Sun, R. A. Tomas, and J. Screen, 2016b: Does ocean coupling matter for the northern extratropical response to projected Arctic sea ice loss? Geophys. Res. Lett., 43, 2149–2157, https://doi.org/10.1002/2016GL067792.
Di Luca, A., R. de Elía, M. Bador, and D. Argüeso, 2020: Contribution of mean climate to hot temperature extremes for present and future climates. Wea. Climate Extremes, 28, 100255, https://doi.org/10.1016/j.wace.2020.100255.
Eayrs, C., X. Li, M. N. Raphael, and D. M. Holland, 2021: Rapid decline in Antarctic sea ice in recent years hints at future change. Nat. Geosci., 14, 460–464, https://doi.org/10.1038/s41561-021-00768-3.
England, M. R., L. Polvani, and L. Sun, 2018: Contrasting the Antarctic and Arctic atmospheric responses to projected sea ice loss in the late twenty-first century. J. Climate, 31, 6353–6370, https://doi.org/10.1175/JCLI-D-17-0666.1.
England, M. R., L. M. Polvani, and L. Sun, 2020a: Robust Arctic warming caused by projected Antarctic sea ice loss. Environ. Res. Lett., 15, 104005, https://doi.org/10.1088/1748-9326/abaada.
England, M. R., L. M. Polvani, L. Sun, and C. Deser, 2020b: Tropical climate responses to projected Arctic and Antarctic sea-ice loss. Nat. Geosci., 13, 275–281, https://doi.org/10.1038/s41561-020-0546-9.
England, M. R., I. Eisenman, and T. J. W. Wagner, 2022: Spurious climate impacts in coupled sea ice loss simulations. J. Climate, 35, 3801–3811, https://doi.org/10.1175/JCLI-D-21-0647.1.
Fischer, E. M., and R. Knutti, 2015: Anthropogenic contribution to global occurrence of heavy-precipitation and high-temperature extremes. Nat. Climate Change, 5, 560–564, https://doi.org/10.1038/nclimate2617.
Gillett, N. P., T. D. Kell, and P. D. Jones, 2006: Regional climate impacts of the Southern Annular Mode. Geophys. Res. Lett., 33, L23704, https://doi.org/10.1029/2006GL027721.
Gong, D., and S. Wang, 1999: Definition of Antarctic Oscillation index. Geophys. Res. Lett., 26, 459–462, https://doi.org/10.1029/1999GL900003.
Hurrell, J. W., and Coauthors, 2013: The Community Earth System Model: A framework for collaborative research. Bull. Amer. Meteor. Soc., 94, 1339–1360, https://doi.org/10.1175/BAMS-D-12-00121.1.
Jones, J. M., and M. Widmann, 2003: Instrument- and tree-ring-based estimates of the Antarctic Oscillation. J. Climate, 16, 3511–3524, https://doi.org/10.1175/1520-0442(2003)016<3511:IATEOT>2.0.CO;2.
Kay, J. E., M. M. Holland, C. M. Bitz, E. Blanchard-Wrigglesworth, A. Gettelman, A. Conley, and D. Bailey, 2012: The influence of local feedbacks and northward heat transport on the equilibrium Arctic climate response to increased greenhouse gas forcing. J. Climate, 25, 5433–5450, https://doi.org/10.1175/JCLI-D-11-00622.1.
Kay, J. E., and Coauthors, 2015: The Community Earth System Model (CESM) Large Ensemble Project: A community resource for studying climate change in the presence of internal climate variability. Bull. Amer. Meteor. Soc., 96, 1333–1349, https://doi.org/10.1175/bams-d-13-00255.1.
Kharin, V. V., F. W. Zwiers, X. Zhang, and M. Wehner, 2013: Changes in temperature and precipitation extremes in the CMIP5 ensemble. Climatic Change, 119, 345–357, https://doi.org/10.1007/s10584-013-0705-8.
Khokhlov, V. N., A. V. Glushkov, and I. A. Tsenenko, 2004: Atmospheric teleconnection patterns and eddy kinetic energy content: Wavelet analysis. Nonlinear Processes Geophys., 11, 295–301, https://doi.org/10.5194/npg-11-295-2004.
Kidston, J., A. S. Taschetto, D. W. J. Thompson, and M. H. England, 2011: The influence of Southern Hemisphere sea-ice extent on the latitude of the mid-latitude jet stream. Geophys. Res. Lett., 38, L15804, https://doi.org/10.1029/2011GL048056.
Li, C., F. Zwiers, X. Zhang, G. Li, Y. Sun, and M. Wehner, 2021: Changes in annual extremes of daily temperature and precipitation in CMIP6 models. J. Climate, 34, 3441–3460, https://doi.org/10.1175/JCLI-D-19-1013.1.
Lin, L., Z. Wang, Y. Xu, and Q. Fu, 2016: Sensitivity of precipitation extremes to radiative forcing of greenhouse gases and aerosols. Geophys. Res. Lett., 43, 9860–9868, https://doi.org/10.1002/2016GL070869.
Liu, J., J. A. Curry, H. Wang, M. Song, and R. M. Horton, 2012: Impact of declining Arctic sea ice on winter snowfall. Proc. Natl. Acad. Sci. USA, 109, 4074–4079, https://doi.org/10.1073/pnas.1114910109.
Liu, R., H. Su, K.-N. Liou, J. H. Jiang, Y. Gu, S. C. Liu, and C.-J. Shiu, 2018: An assessment of tropospheric water vapor feedback using radiative kernels. J. Geophys. Res. Atmos., 123, 1499–1509, https://doi.org/10.1002/2017JD027512.
Menéndez, C. G., V. Serafini, and H. Le Treut, 1999: The effect of sea-ice on the transient atmospheric eddies of the Southern Hemisphere. Climate Dyn., 15, 659–671, https://doi.org/10.1007/s003820050308.
Overland, J., and Coauthors, 2019: The urgency of Arctic change. Polar Sci., 21, 6–13, https://doi.org/10.1016/j.polar.2018.11.008.
Parkinson, C. L., 2019: A 40-y record reveals gradual Antarctic sea ice increases followed by decreases at rates far exceeding the rates seen in the Arctic. Proc. Natl. Acad. Sci. USA, 116, 14 414–14 423, https://doi.org/10.1073/pnas.1906556116.
Peings, Y., and G. Magnusdottir, 2014: Response of the wintertime Northern Hemisphere atmospheric circulation to current and projected Arctic sea ice decline: A numerical study with CAM5. J. Climate, 27, 244–264, https://doi.org/10.1175/JCLI-D-13-00272.1.
Pezza, A. B., T. Durrant, I. Simmonds, and I. Smith, 2008: Southern Hemisphere synoptic behavior in extreme phases of SAM, ENSO, sea ice extent, and southern Australia rainfall. J. Climate, 21, 5566–5584, https://doi.org/10.1175/2008JCLI2128.1.
Raphael, M. N., W. Hobbs, and I. Wainer, 2011: The effect of Antarctic sea ice on the Southern Hemisphere atmosphere during the southern summer. Climate Dyn., 36, 1403–1417, https://doi.org/10.1007/s00382-010-0892-1.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.
Riahi, K., and Coauthors, 2017: The shared socioeconomic pathways and their energy, land use, and greenhouse gas emissions implications: An overview. Global Environ. Change, 42, 153–168, https://doi.org/10.1016/j.gloenvcha.2016.05.009.
Roach, L. A., and Coauthors, 2020: Antarctic sea ice area in CMIP6. Geophys. Res. Lett., 47, e2019GL086729, https://doi.org/10.1029/2019GL086729.
Screen, J. A., C. Deser, and L. Sun, 2015: Projected changes in regional climate extremes arising from Arctic sea ice loss. Environ. Res. Lett., 10, 084006, https://doi.org/10.1088/1748-9326/10/8/084006.
Seneviratne, S. I., and Coauthors, 2012: Changes in climate extremes and their impacts on the natural physical environment. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation, C. B. Field, Ed., Cambridge University Press, 109–230.
Seneviratne, S. I., and Coauthors, 2021: Weather and climate extreme events in a changing climate. Climate Change 2021: The Physical Science Basis, V. Masson-Delmotte, Ed., Cambridge University Press, 1513–1766.
Shu, Q., Z. Song, and F. Qiao, 2015: Assessment of sea ice simulations in the CMIP5 models. Cryosphere, 9, 399–409, https://doi.org/10.5194/tc-9-399-2015.
Sillmann, J., V. V. Kharin, F. W. Zwiers, X. Zhang, and D. Bronaugh, 2013: Climate extremes indices in the CMIP5 multimodel ensemble: Part 2. Future climate projections. J. Geophys. Res. Atmos., 118, 2473–2493, https://doi.org/10.1002/jgrd.50188.
Sillmann, J., and Coauthors, 2019: Extreme wet and dry conditions affected differently by greenhouse gases and aerosols. npj Climate Atmos. Sci., 2, 24, https://doi.org/10.1038/s41612-019-0079-3.
Smith, D. M., N. J. Dunstone, A. A. Scaife, E. K. Fiedler, D. Copsey, and S. C. Hardiman, 2017: Atmospheric response to Arctic and Antarctic sea ice: The importance of ocean–atmosphere coupling and the background state. J. Climate, 30, 4547–4565, https://doi.org/10.1175/JCLI-D-16-0564.1.
Song, M.-R., S.-Y. Wang, Z. Zhu, and J.-P. Liu, 2021: Nonlinear changes in cold spell and heat wave arising from Arctic sea-ice loss. Adv. Climate Change Res., 12, 553–562, https://doi.org/10.1016/j.accre.2021.08.003.
Sun, L., C. Deser, R. A. Tomas, and M. Alexander, 2020: Global coupled climate response to polar sea ice loss: Evaluating the effectiveness of different ice-constraining approaches. Geophys. Res. Lett., 47, e2019GL085788, https://doi.org/10.1029/2019GL085788.
Terray, L., 2021: A dynamical adjustment perspective on extreme event attribution. Wea. Climate Dyn., 2, 971–989, https://doi.org/10.5194/wcd-2-971-2021.
Turner, J., T. J. Bracegirdle, T. Phillips, G. J. Marshall, and J. S. Hosking, 2013: An initial assessment of Antarctic sea ice extent in the CMIP5 models. J. Climate, 26, 1473–1484, https://doi.org/10.1175/JCLI-D-12-00068.1.
Turner, J., J. S. Hosking, T. J. Bracegirdle, G. J. Marshall, and T. Phillips, 2015: Recent changes in Antarctic sea ice. Philos. Trans. Roy. Soc., A373, 20140163, https://doi.org/10.1098/rsta.2014.0163.
Turner, J., T. Phillips, G. J. Marshall, J. S. Hosking, J. O. Pope, T. J. Bracegirdle, and P. Deb, 2017: Unprecedented springtime retreat of Antarctic sea ice in 2016. Geophys. Res. Lett., 44, 6868–6875, https://doi.org/10.1002/2017GL073656.
Westra, S., L. V. Alexander, and F. W. Zwiers, 2013: Global increasing trends in annual maximum daily precipitation. J. Climate, 26, 3904–3918, https://doi.org/10.1175/JCLI-D-12-00502.1.
Yu, L. J., Z. Zhang, M. Zhou, S. Zhong, D. Lenschow, H. Hsu, H. Wu, and B. Sun, 2012: Influence of the Antarctic Oscillation, the Pacific–South American modes and the El Niño–Southern Oscillation on the Antarctic surface temperature and pressure variations. Antarct. Sci., 24, 59–76, https://doi.org/10.1017/S095410201100054X.
Zhang, X., L. Alexander, G. C. Hegerl, P. Jones, A. K. Tank, T. C. Peterson, B. Trewin, and F. W. Zwiers, 2011: Indices for monitoring changes in extremes based on daily temperature and precipitation data. Wiley Interdiscip. Rev.: Climate Change, 2, 851–870, https://doi.org/10.1002/wcc.147.
Zheng, F., J. Li, L. Wang, F. Xie, and X. Li, 2015: Cross-seasonal influence of the December–February Southern Hemisphere Annular Mode on March–May meridional circulation and precipitation. J. Climate, 28, 6859–6881, https://doi.org/10.1175/JCLI-D-14-00515.1.