1. Introduction
The Arctic has warmed almost 4 times faster than the global average during recent decades (Rantanen et al. 2022), a phenomenon known as Arctic amplification (AA) (Screen and Simmonds 2010; England et al. 2021). Climate projections suggest AA will continue in the future if greenhouse gas concentrations continue to rise (Notz and Stroeve 2016; Dai et al. 2019). Such rapid warming has been associated with the loss of roughly half of the summer Arctic sea ice cover in a little over 40 years (Wang and Overland 2009; Stroeve and Notz 2018). The latest phase 6 of the Coupled Model Intercomparison Project (CMIP6) projections show that there will be an ice-free Arctic in summer for the first time before 2050 in all scenarios and in almost all models (Notz and SIMIP Community 2020).
Arctic sea ice loss has had, and is projected to continue to have, severe impacts on Arctic weather and climate (e.g., Landrum and Holland 2020). Furthermore, Arctic sea ice loss has been linked to the changes in weather and climate far outside the Arctic. For example, it has been connected to more frequent cold wave events over midlatitudes (Vihma 2014; Wu et al. 2017; Cohen et al. 2020, 2021; Xu et al. 2021). However, other studies suggest that the Arctic sea ice loss has had a minimal effect on midlatitude cold winters (Blackport et al. 2019), or even reduces the risk of cold extremes (Blackport et al. 2022).
The potential influence of Arctic sea ice loss on precipitation has been less studied than that on temperature. It is well known that sea ice loss causes an increase in winter-mean precipitation over high latitudes (Screen and Simmonds 2010; McCrystall et al. 2021), associated with the increasing upward heat and moisture flux from the open ocean (Bintanja and Selten 2014; Deser et al. 2016; Levine et al. 2021). However, the precipitation response to sea ice loss may not be confined to high latitudes. Remote impacts of Arctic sea ice loss may include wintertime drying in California (Cvijanovic et al. 2017) and increased Sahel precipitation (Monerie et al. 2019). Arctic sea ice loss may also cause more frequent strong EI Niño events (Liu et al. 2022), with consequent effects on global precipitation patterns.
Arctic sea ice loss may affect precipitation extremes as well as seasonal averages. Screen et al. (2015) found that Arctic sea ice loss will increase the likelihood and severity of wet extremes over high latitudes while decreasing the likelihood of dry days over midlatitude Eurasia. Liu et al. (2021) found a high correlation between AA and extreme precipitation in models and observations over the Northern Hemisphere. Bailey et al. (2021) linked the Arctic sea ice loss with increased evaporation and European extreme snowfall and suggested that abnormally low Barents sea ice cover contributed to an extreme snowfall event across Europe in 2018.
There exists large disagreement between different modeling studies on the midlatitude effects of Arctic sea ice loss (Screen et al. 2018). For example, some previous studies found a negative phase of North Atlantic Oscillation (NAO) response to Arctic sea ice loss (Kim et al. 2014; Mori et al. 2014), while others found a positive NAO phase response (Cassano et al. 2014; Screen et al. 2014). Such inconsistency may be due to, for example, different model physics (Sun et al. 2015), different magnitudes and spatial patterns of imposed sea ice loss (McKenna et al. 2018; Zhang and Screen 2021), different experimental approaches (Tomas et al. 2016; Oudar et al. 2017; Smith et al. 2017), too-weak eddy feedback (Smith et al. 2022; Hardiman et al. 2022; Screen et al. 2022), and difficulties separating the forced response from large internal variability (Liang et al. 2020; Peings et al. 2021). To minimize these challenges, here we use large ensembles of coordinated experiments with multiple models, produced by the Polar Amplification Model Intercomparison Project (PAMIP).
While there is scientific merit in quantifying and understanding the response to sea ice loss, it is important to remember that, in the real world, sea ice loss does not occur in isolation from other aspects of global climate change. It has been suggested that ocean warming, especially in the tropics, has an opposite effect on the midlatitude circulation to that from Arctic sea ice loss (Barnes and Screen 2015). Lower-tropospheric warming in the Arctic due to Arctic sea ice loss favors a more equatorward jet, whereas tropical upper-tropospheric warming due to sea surface temperature (SST) change favors a more poleward jet; this is referred as the “tug of war” (Barnes and Screen 2015; Screen et al. 2018). Although the tug-of-war is often mentioned in the context of the latitudinal shift of the jet stream, no studies to our knowledge have considered the potentially competing effects of Arctic sea ice loss and ocean warming on midlatitude precipitation variability and extremes. Thus, one main aim of this study is to directly compare the effects of these two forcings on precipitation.
Here, we use large ensembles from PAMIP and CMIP6 to study the impacts of future Arctic sea ice loss and global ocean warming on wintertime precipitation. We seek to answer three main questions:
-
What is the response of precipitation and extreme precipitation events to Arctic sea ice loss, and what are the dominant physical mechanisms?
-
How does the precipitation response to sea ice loss and ocean warming compare in magnitude and spatial pattern?
-
Can the linear combination of effects of global ocean warming and Arctic sea ice loss explain projected precipitation changes at 2°C of global warming?
2. Data and methods
a. Models and experiments
To study the response to Arctic sea ice loss and global ocean warming separately, we used experiments from eight models that contributed to PAMIP (Smith et al. 2019). The models and experiments used are shown in Table 1. In summary, these experiments are run with prescribed present-day sea ice concentration (SIC) and present-day SST (pdSST-pdSIC); future Arctic SIC and future Arctic SST (where sea ice is reduced) but present-day SST outside the Arctic (pdSST-futArcSIC); and present-day SIC and future SST (futSST-pdSIC). Using such combinations of SST and SIC, we can study the response to each forcing separately. Figure 1 shows the prescribed SST and SIC forcings fields for the experiments. Further details on the SIC and SST forcing can be found in appendix A of Smith et al. (2019). To compare the responses to surface boundary changes, as seen in PAMIP (Fig. 1), with climate change projections under increasing greenhouse gases, we sampled output from the CMIP6 (Eyring et al. 2016) historical and shared socioeconomic pathway (SSP) experiments (ScenarioMIP) from the same eight models (Table 1). Although PAMIP is also a part of CMIP6, to distinguish the two different sets of experiments, here we refer to the CMIP6 historical and ScenarioMIP as “CMIP6.” We identified the time period in each model, experiment, and ensemble member when the 30-yr-mean global near-surface temperature first reaches 14.24° and 15.67°C, following the definitions of “present day” and “future” climates in the PAMIP (Smith et al. 2019). Cold biases in some models (marked by − in Table 1) means they do not reach the present-day threshold by 2014 (i.e., in the historical run), while warm biases in other models (marked by + in Table 1) means the future threshold is passed before 2014. In such cases, we sample from solely the historical or future simulations, respectively. In other cases, the thresholds are reached in a 30-yr window overlapping the historical and future simulations. Here, we concatenated the historical and future simulations, picking the same ensemble members (i.e., the corresponding future simulations branched from each historical run) before sampling. We sampled all available ensemble members and for the projections, all available scenarios. For CESM2, we supplemented the core CMIP6 simulations with those from the CESM2 large ensemble (CESM2-LENS; Rodgers et al. 2021), performed to the same experimental protocol. The sampling of CMIP6 is effective at minimizing the differences in sea ice and SST between PAMIP and CMIP6. However, some differences still remain, which are further discussed in section 3.
The details of models, experiments, and the number of members used in this study. The models that only have monthly precipitation data are marked with a caret (∧). Models that have 6-hourly sea level pressure data (used for cyclone tracking) and daily geopotential height, zonal wind, air temperature data (used to calculate Eady growth rate) are marked with an asterisk (*). Models for which the “present day” period is sampled from their projections (not their historical runs) are marked by a minus sign (−), while models for which the “future” period is sampled from the historical runs (not projections) are marked by a plus sign (+).
(a) Winter sea ice concentration (%) changes, future minus present day, prescribed in the PAMIP futArcSIC and (b) that for sea surface temperature (°C) changes in the PAMIP futArcSIC and (c) futSST experiments.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
b. Observational data
The ERA5 dataset (Hersbach et al. 2020), NCEP Climate Forecast System Reanalysis (CFSR; Saha et al. 2010), Modern-Era Retrospective Analysis for Research and Applications (MERRA; Rienecker et al. 2011), and Global Precipitation Climatology Project (GPCP; Adler et al. 2018) are all used in our study to evaluate the simulated present-day winter precipitation (pdSST-pdSIC).
c. Response estimation and statistical significance
Throughout the study, analysis will be performed for the Northern Hemisphere winter defined as December–February. As well as gridded fields, we consider two regional domains: the North Atlantic and midlatitude Atlantic (locations are shown in Fig. 3). All models have been regridded to the lowest horizontal resolution among these models (CanESM5, ∼2.81° × 2.81°) using bilinear interpolation. Responses are calculated as the ensemble-mean difference between the perturbation (pdSST-futArcSIC or futSST-pdSIC) and control experiment (pdSST-pdSIC). We average the members across the models before calculating the multimodel ensemble-mean, so each model is given equal weight. We used the two-tailed Student’s t test at the 95% confidence level to assess statistical significance.
d. Extreme indices
According to the framework of the Expert Team on Climate Change Detection and Indices (ETCCDI; Zhang et al. 2011), six extreme dry and wet indices are used to describe the wintertime extreme events. The definitions are shown in Table 2.
Definitions of extreme indices.
e. Cyclone tracking
The cyclone feature tracking method is following Hoskins and Hodges (2002), using 6-hourly sea level pressure (SLP) data from PAMIP experiments. SLP minima are identified in the gridded data (truncated to T63 resolution with wavenumbers smaller than or equal to 5 filtered out to eliminate planetary scales) and defined as the cyclone centers. Centers are joined up using a nearest-neighbor approach, before the tracks are optimized by minimizing a cost function (Hodges 1994, 1999). Only storms that last more than 48 h and travel at least 1000 km are retained. Track density (the number of tracks passing through a grid box at a particular time) and seasonal-mean intensity (mean SLP intensity after removal of background field of all tracks passing through a grid box) are two characteristics we used to describe the cyclone changes under future Arctic sea ice loss and ocean warming.
f. Eady growth rate
3. Results
a. Precipitation response to future Arctic sea ice loss
First, we compare the observed/reanalysis and simulated present-day (pdSST-pdSIC) precipitation in PAMIP (Fig. 2). The multimodel-mean-simulated winter precipitation in the North Atlantic (129 mm month−1) and midlatitude Atlantic (88 mm month−1) are close to the average of the four observational products (135 and 82 mm month−1, respectively). The simulated values lie within the observational uncertainty, as estimated by the spread of the four observation-based datasets. Although individual models have larger biases than the multimodel-mean, we found that there is no statistically significant relationship between the magnitudes of the individual model biases and their responses to sea ice loss (not shown). We interpret internal variability to be a larger source of uncertainty in the diagnosed responses than mean model biases.
Comparison between observational data (1979–2008) and PAMIP present-day winter precipitation over the North Atlantic (red bars) and midlatitude Atlantic (blue bars).
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
The wintertime precipitation response to Arctic sea ice loss in each model is shown in Figs. 3a–h. Precipitation increases over the regions of Arctic sea ice loss, for example, the Barents–Kara Sea, Hudson Bay, and the Sea of Okhotsk, in all models. This increase is likely related to the local thermodynamic effects of sea ice loss (Deser et al. 2016; Levine et al. 2021). Outside the Arctic, the most notable response is a drying signal over the North Atlantic, albeit with variation in location and magnitude across the models. The multimodel ensemble-mean (Fig. 3i) shows robust drying over the North Atlantic (NA; Fig. 3i, red box), being statistically significant in the full ensemble and with most models agreeing on the sign. In the ensemble-mean and several models, there is a significant wetting response over the midlatitude Atlantic (MA; Fig. 3i, black box). However, the wetting response over MA (0–5 mm month−1) is smaller than the drying response over NA (5–10 mm month−1). We also note that there are some significant wetting signals over the North Pacific and drying signal over Siberia, the Ural region, and the U.S. East Coast. Nevertheless, the response in the above regions is weak (0–5 mm month−1) and is not robust across the models.
(a)–(h) Winter-mean precipitation (mm month−1) response to future Arctic sea ice loss across models and (i) for the multimodel-mean. Hatching denotes a statistically significant response at the 95% confidence level. Stippling in (i) denotes at least seven individual models agree on the sign of the change. The red and black boxed areas are defined as the North Atlantic (NA) and midlatitude Atlantic (MA), respectively.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
Next, we consider the response of winter daily precipitation extremes to future Arctic sea ice loss (Fig. 4, see “methods” for detailed definition). The simple precipitation intensity index [the simple daily intensity index (SDII)] reduces by 0.3 mm day−1 over NA (Fig. 4a), consistent with the pattern of the winter-mean precipitation response (Fig. 3i). Precipitation intensity also significantly decreases by more than 0.3 mm day−1 over the Chukchi and Bering Seas and Hudson Bay, which is in contrast to the winter-mean wetting signal in these regions. On the contrary, there is also an increase of precipitation intensity along the U.S. East Coast and midlatitude Atlantic; the latter is consistent with the seasonal-mean wetting signal here (Fig. 3i). The number of dry days (R1mm; Fig. 4b) and dry-spell duration (CDD; Fig. 4c) increase significantly over NA and extending to Greenland, while decreasing significantly over the Arctic and subarctic (in the vicinity of sea ice loss) and MA. The number of wet days and very wet days (Figs. 4d,e) decreases significantly over NA while increasing significantly over MA. Wet-spell duration (CWD) increases significantly over the Arctic and subarctic seas. However, there are no significant increases in wet and very wet days in these regions, which might indicate that the wetting signal here is caused by the increasing number of light or moderate rainy days rather than extreme rainy days. CWD increases significantly over MA and decreases significantly over NA, which aligns with the changes of dry and wet days in these regions.
Multimodel ensemble-mean (a)–(c) dry and (d)–(f) wet extreme indices response to Arctic sea ice loss. (a) Simple precipitation intensity index (SDII); (b) number of dry days (R1mm); (c) maximum length of dry spell (CDD); (d) number of wet days (R10mm)); (e) number of very wet days (R20mm)); (f) maximum length of wet spell (CWD). The color bars (b) and (c) are reversed, so that in all panels, brown colors denote a drying response while blue denote a wetting response. The hatching denotes the significance at the 95% confidence level, while the stippling denotes six out of seven models agree with the sign.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
The winter-mean drying over NA (Fig. 3i) is related to decreased daily precipitation intensity, more dry days, longer dry spells, fewer wet days, and shorter wet spells (Fig. 4), whereas the winter-mean wetting over MA (Fig. 3i) is associated with increased precipitation intensity, fewer dry days, shorter dry spells, more wet days, and longer wet spells. The precipitation change in these two regions is illustrated further in Fig. 5, which shows the probability distribution functions (PDF) of daily precipitation in the control experiment (black line) and future Arctic sea ice loss experiment (red line). While sea ice loss leads to more dry days and fewer wet days, the change in the PDF of daily precipitation is subtle compared to the relatively larger changes in response to SST, which we will return to later.
Probability distribution functions of daily precipitation (mm day−1) for present-day (black line), future Arctic Sea ice (red), and future SST (blue) experiments for the (a) North Atlantic and (b) midlatitude Atlantic. The gray-shaded bands denote the range of observed estimates from the four observation-based products.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
b. Response of precipitation to future global ocean warming
We next study the impacts of future SST on the Northern Hemisphere winter-mean precipitation across models (Fig. 6). The response to ocean warming is more robust between models than is the response to sea ice loss. All models show a significant wetting signal over high latitudes and much of the midlatitudes (Fig. 6i), which may reflect the first-order thermodynamical response to warming. However, regions of significant drying are also simulated, over southern Europe and the Mediterranean. Over the Atlantic–Europe sector, there is a strong dipole—wetting over NA and drying over MA—that might indicate a shift of the Atlantic storm track. A similar dipole is evident over the Pacific storm-track region.
(a)–(h) Winter-mean precipitation (mm month−1) response to future global ocean warming across models and (i) for the multimodel-mean. Hatching denotes a statistically significant response at the 95% confidence level. Stippling in (i) denotes at least seven of individual models agree on the sign of the change. The red and black boxed areas are defined as the North Atlantic (NA), and midlatitude Atlantic (MA), respectively.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
Comparing the multimodel-mean precipitation response to Arctic sea ice loss (Fig. 3i) and ocean warming (Fig. 6i), we find the responses are opposite to each other in the North Atlantic sector. More specifically, focusing on the two highlighted boxes, we find ocean warming causes wetting over NA and drying over MA, compared to drying and wetting, respectively, in response to sea ice loss. Thus, there appears to be a tug-of-war between Arctic sea ice loss and ocean warming on the North Atlantic wintertime precipitation. In both regions, the response to SST (Fig. 6i) is of a larger magnitude compared to the response to Arctic sea ice loss (Fig. 3i), indicating that the response to SST may overwhelm the effects of Arctic sea ice loss on precipitation over the Atlantic–Europe sector, a point we will return to later in the context of future projections.
Changes in extreme precipitation in response to SST typically follow the changes in the winter-mean (Fig. 7). The dominant wetting signal is accompanied by more intense rainfall (Fig. 7a), fewer dry days (Fig. 7b), shorter dry spells (Fig. 7c), more wet days (Figs. 7d,e), and longer wet spells (Fig. 7f), and vice versa for the drying signals over the Mediterranean and central Pacific (Fig. 7). The dipole over the Atlantic–Europe sector is seen in all extreme indices, with perhaps the exception of wet-spell duration. For most regions and indices, the changes are larger in magnitude for the response to SST than for the response to sea ice loss. For example, precipitation intensity increases by over 0.3 mm day−1 in the North Atlantic and North Pacific sectors, which is larger in magnitude than the decreased precipitation intensity (0.1–0.3 mm) caused by Arctic sea ice loss (Fig. 4a). The larger response to SST than sea ice loss is also clear, returning to Fig. 5, which shows bigger changes in the PDFs of daily precipitation for future ocean warming than for future Arctic sea ice loss.
(a)–(c) Multimodel ensemble-mean dry and (d)–(f) wet extreme indices response to future global ocean warming. (a) Simple precipitation intensity index (SDII); (b) number of dry days (R1 mm); (c) maximum length of dry spell (CDD); (d) number of wet days (R10mm)); (e) number of very wet days (R20mm); (f) maximum length of wet spell (CWD). The color bars (b) and (c) are reversed, so that in all panels, brown colors denote a drying response, while blue denote a wetting response. The hatching denotes the significance at the 95% confidence level, while the stippling denotes six out of seven models agree with the sign.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
c. Combined response to sea ice and SST and comparison with projections
Previous studies have found that the effects of ocean warming and Arctic sea ice loss are linearly additive (Blackport and Kushner 2017; McCusker et al. 2017; Hay et al. 2022). The PAMIP experiments provide further evidence to support additivity (see Fig. A1 in the appendix). Therefore, we add the responses to ocean warming and sea ice loss to get an estimate of their combined effect (Fig. 8a). The sum of the two precipitation responses looks similar to the response to future ocean warming (Fig. 6i), reaffirming that in most regions, the response to SST overwhelms the response to sea ice loss. In high latitudes and in the vicinity of regions of sea ice loss (e.g., Hudson–Baffin Bays and Barents–Kara Sea), the combined wetting response is larger than either of the separate responses, which are of the same sign.
Comparison of precipitation response in PAMIP (futSST + futArcSIC) and CMIP6 (future minus present day) (unit: mm month−1). (a) Sum of precipitation response to futSST and futArcSIC. (b) Precipitation difference between future and present-day period in CMIP6. (c) (a) minus (b). Shading refers to the multimodel ensemble-mean results, and hatching denotes the significance level at 95% confidence level. The red and black boxed areas are defined as the North Atlantic (NA) and midlatitude Atlantic (MA), respectively.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
The analysis so far is based on atmosphere-only experiments. However, due to the lack of ocean–atmosphere coupling, the impacts of Arctic sea ice loss and/or SST may be biased (Deser et al. 2015, 2016). Additionally, besides SST and sea ice changes, direct radiative forcing from greenhouse gases or anthropogenic aerosols may also affect precipitation (Wu et al. 2013; Christidis and Stott 2022). Next, we explore how well the sum of PAMIP experiments match the projected precipitation changes in coupled models forced by scenarios of future radiative forcing. The projected precipitation changes in CMIP6 (Fig. 8b) match well the combined response pattern of PAMIP. The main spatial features of the PAMIP responses already described are also seen in the CMIP6 projections. This suggests that it is reasonable to view the projections as a combined response to SST and sea ice loss, at least in terms of the spatial pattern. There are differences in the magnitudes of the combined PAMIP response and CMIP6 projections. Although these could arise for several reasons, such as a lack of ocean coupling and direct radiative forcing in the PAMIP experiments, we note there is reasonable spatial agreement between the PAMIP and CMIP6 precipitation response difference (Fig. 8c) and differences between the prescribed (in PAMIP) and simulated (in CMIP6) SSTs (Fig. 9c). For example, the excessive wetting in the North Atlantic and North Pacific in PAMIP compared to CMIP6 is collocated with larger ocean warming in PAMIP than in CMIP6 (Fig. 9c). So, at least to some extent, the differences between PAMIP and CMIP6 precipitation response likely stem from the PAMIP boundary forcings not perfectly matching the projected SST and sea ice changes in CMIP6.
Comparison of (a) winter sea surface temperature (°C) changes (future minus present day) prescribed in the PAMIP experiments and (b) that simulated by the CMIP6 coupled experiments at comparable global warming levels, and (c) their differences [(a) minus (b)]. (d)–(f) As in (a)–(c), but for sea ice concentration (%).
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
Figure 10 summarizes the winter precipitation responses to Arctic sea ice loss, global ocean warming, their sum, and the CMIP6 projection for NA (Fig. 10a) and MA (Fig. 10b), for all models. Most models agree on a relatively weak drying of the NA region in response to sea ice loss and a relatively strong wetting in response to ocean warming. Hence, the wetting response dominates, which is supported by most CMIP6 models. In the MA region, however, the two responses in the PAMIP experiments cancel each other out, as the wetting in response to sea ice loss and the drying in response to ocean warming are of similar strength. Five out of eight models show no significant precipitation change over this region.
The regional-mean precipitation response to future Arctic sea ice loss and global ocean warming (unit: mm month−1). Precipitation response to sea ice loss (black circles), global ocean warming (blue circles), and their sum (green circles) for each model. Magenta circles show the precipitation changes simulated in CMIP6 experiments. Statistically significant (95% confidence) individual model responses are shown by filled circles; otherwise, circles are unfilled.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
d. Mechanisms of opposing influences over North Atlantic
Finally, we consider the mechanisms driving the contrasting precipitation responses to Arctic sea ice loss and global ocean warming over the North Atlantic sector. As alluded to earlier, a likely explanation for the dipole precipitation response is a shift in the Atlantic storm track. Arctic sea ice loss may weaken the meridional temperature gradient and then, through eddy–mean interactions, the storm tracks may weaken and shift equatorward (Smith et al. 2022). On the contrary, ocean warming preferentially warms the lower latitudes, increasing the meridional temperature gradient, and may cause an intensification and poleward shift of storm tracks (Graff and LaCasce 2012). A shift of storm tracks will directly influence precipitation and extremes especially in midlatitudes (Lehmann and Coumou 2015; Shaw et al. 2016).
Figures 11a and 11b shows the multimodel ensemble-mean storm-track density (number of tracks that pass through a grid box per month) and intensity (mean sea level pressure intensity after removal of background field of all tracks passing through a grid box) in the present-day experiment. The North Atlantic storm track is clear as a region of frequent and intense storms. There is a southward shift of the storm track in response to Arctic sea ice loss (Fig. 11d), which is consistent with Smith et al. (2022). The most prominent area of decreased storms is the NA, where the maximum drying response is simulated. (Fig. 11f). Figure 11e shows that there is also decreasing intensity of storms in the NA. Fewer and weaker storms passing through NA provide a physical explanation for the drying in response to sea ice loss. Conversely, there is increased storm density close to MA, where we found a wetting response. We note the wetting response is located slightly south of the region of increased storm density response (Fig. 10d), possibly because the maximum precipitation occurs on the southern edge of the storm track, associated with trailing cold fronts, which can be seen in Fig. 11c and is also consistent with previous literature (Fig. 2 in Booth et al. 2018; Fig. 2 in Kodama et al. 2019).
Multimodel ensemble-mean (left) storm density (unit: number month−1), (center) intensity (unit: hPa), and (right) precipitation (unit: mm month−1) in (top) PAMIP present-day experiment, (middle) response to future Arctic sea ice loss, and (bottom) future global ocean warming. The hatching denotes the significance at the 95% confidence level. The green contours in (d) and (g) indicate the baseline climatology (contour value: 28 month−1).
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
Figure 11g shows a northeastern extension of storm tracks over the Atlantic in response to future global ocean warming, while the intensity of storms significantly increases in NA (Fig. 10h). The density and intensity of Mediterranean storms significantly decrease over MA, which could be linked to the drying signal there (Fig. 6i). Similar to the competing effect of future Arctic sea ice loss and ocean warming on precipitation, we also find the tug-of-war on the North Atlantic storm tracks. The above results support the important, but opposite, role of extratropical cyclones in the precipitation responses to future Arctic sea ice loss and ocean warming.
Since the existence of cyclones and storm tracks is dependent on baroclinic instability, we use the EGR to measure the responses of baroclinicity to future Arctic sea ice loss and global ocean warming (Fig. 12). There is a dipole pattern of baroclinicity over the North Atlantic in response to future Arctic sea ice loss (Fig. 12d), which aligns well with the track density response (Fig. 11d) and is also consistent with previous literature (Figs. 13e,f in Oudar et al. 2017). With regard to the EGR in response to future ocean warming (Fig. 12g), there is a significantly increasing EGR over NA which also matches the track density and mean intensity response here (Figs. 11g,h). The changes of EGR are determined by the changes in static stability and changes in vertical wind shear (see methods section). According to their definition, the vertical wind shear is closely related to the meridional temperature gradient (via the thermal wind relation), and the Brunt–Väisälä frequency is related to the vertical gradient of potential temperature (or static stability). Comparing the vertical wind shear response to future Arctic sea ice loss (Fig. 12e) and future ocean warming (Fig. 12h), we find they look similar to the EGR responses, which indicates that the response of EGR is dominated by the changes in vertical wind shear. The responses of Brunt–Väisälä frequency are shown in Figs. 12f and 12i. We find negative anomalies in the high latitudes, which is due to a decrease in vertical temperature gradient and atmospheric stability caused by the increase in upward heat flux warming the lower atmosphere under the conditions of Arctic sea ice loss and ocean warming (Oudar et al. 2017; Mioduszewski et al. 2018). In summary, the EGR and vertical wind shear responses further support the mechanisms whereby Arctic warming causes a reduction in the meridional temperature gradient and a southward shift in both baroclinic instability and storm tracks over the North Atlantic (Smith et al. 2022).
Maximum Eady growth rate and related two components response to future Arctic sea ice loss and future ocean warming (unit: day−1). (a),(d),(g) Maximum Eady growth rate, (b),(e),(h) vertical wind shear, and (c),(f),(i) Brunt–Väisälä frequency in (top) PAMIP present-day experiment, (middle) response to future Arctic sea ice loss, and (bottom) future ocean warming. The hatching denotes the significance at the 95% confidence level.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
4. Conclusions
We have used large ensembles from PAMIP and CMIP6 to examine the different responses of winter precipitation and extreme events to Arctic sea ice loss and global ocean warming. The models robustly indicate that NA becomes significantly drier in response to future Arctic sea ice loss, associated with decreasing precipitation intensity as well as an increasing number of dry days. Conversely, MA becomes wetter in response to future sea ice loss. In both these regions, future global ocean warming is simulated to have stronger, but opposite, effects on wintertime precipitation and extremes. The linear combination of precipitation response to future Arctic sea ice loss and global ocean warming shows that the latter will completely cancel out the influence of the former over NA and MA. Comparing with the precipitation change at 2°C global warming projected in CMIP6, the combined response in PAMIP matches well the spatial pattern, albeit with differences in magnitude that appear to relate to differences between the SST prescribed in PAMIP and that simulated in CMIP6. The precipitation responses are closely related to changes in storms (Chang et al. 2022). More specifically, there is an equatorward shift of the storm tracks in response to future Arctic sea ice loss and a northeastern extension of the North Atlantic storm tracks and weakening of Mediterranean storm tracks in response to ocean warming. These storm-track changes provide a dynamical mechanism for opposite precipitation responses to future Arctic sea ice loss and ocean warming over NA and MA. In the CMIP6 projections, Priestley and Catto (2022) found a poleward shift of the North Pacific storm tracks, an extension of North Atlantic storm tracks to Europe, and a weakening of Mediterranean storm tracks in winter. Such changes are very similar to the response to global ocean warming in PAMIP (Fig. 10g). Thus, global ocean warming is the dominant forcing (at least relative to sea ice loss) of changes in future storm tracks.
We close by considering some caveats and directions for future research. We have only shown results for winter. In other seasons, the precipitation response to Arctic sea ice loss is weaker and the response to global ocean warming is much more dominant. The identified responses to sea ice loss are small compared to internal variability. For example, the multimodel ensemble-mean precipitation response to Arctic sea ice loss is ∼3% of the seasonal climatological-mean precipitation in the present-day experiment over NA (∼90 mm month−1). For comparison, the NA precipitation response to SST is ∼10% of the climatological mean. However, Smith et al. (2022) suggested that the storm-track response to Arctic sea ice loss is too weak in models compared to the real world because of too-weak eddy feedback. This opens the possibility that the precipitation response to sea ice loss could also be underestimated. We also note that CMIP6 historical simulations underestimate Arctic sea ice loss compared to observations (Shu et al. 2020; Shen et al. 2021). Therefore, 2°C global warming might be associated with greater Arctic sea ice loss than simulated in CMIP6 (Fig. 1e). Both or either too little sea ice loss or a too-weak sensitivity to sea ice loss would lead to an underestimation of the projected precipitation change in response to sea ice loss. Future work could seek to observationally constrain the precipitation response to sea ice loss. Our quantitative comparison of the PAMIP and CMIP6 experiments is hampered by differences between the prescribed and simulated sea ice and SST. A cleaner decomposition would require PAMIP-like simulations but with model-specific sea ice and SST states prescribed, taken from the CMIP6 projections of each respective model, which is left for future work.
Acknowledgments.
We thank the modeling groups that contributed to the PAMIP and CMIP6. We thank Rym Msadek, Lantao Sun, Michael Sigmond, Rosie Eade, and Hsin-Chien Liang for providing us with additional PAMIP output. We thank David Stephenson, William Seviour, and Matthew Priestley for their discussion and suggestions. H.Y. and M.X were supported by the China Scholarship Council. J.A.S, S.H., and J.L.C were funded by Natural Environment Research Council Grant NE/V005855/1.
Data availability statement.
The PAMIP and CMIP6 data are accessible on the Earth System Grid Federation website (https://esgf-node.llnl.gov/search/cmip6/). CESM2-LENS data are accessible on https://www.cesm.ucar.edu/projects/community-projects/LENS2/.
APPENDIX
Verification of Linear Additivity
The extra PAMIP experiments and models used. These simulations are run with prescribed pre-industrial sea ice concentration (SIC) and sea surface temperature (SST) (piSST-piSIC); pre-industrial Arctic SIC and Arctic SST (where sea ice is reduced) but present-day SST outside the Arctic (pdSST-piArcSIC); and present-day SIC and pre-industrial SST (piSST-pdSIC). The last three columns indicate the number of each model and each experiment.
Verification of linear additive of sea ice (SIC) and sea surface temperature (SST) response (unit: mm month−1). (a) Sum of precipitation response to present-day conditions’ SST and SIC. (b) Direct precipitation response to present-day SST and SIC. (c), (b) minus (a). Shading refers to the multimodel ensemble-mean results, and hatching denotes the significance level at 95% confidence level.
Citation: Journal of Climate 36, 15; 10.1175/JCLI-D-22-0774.1
REFERENCES
Adler, R. F., and Coauthors, 2018: The Global Precipitation Climatology Project (GPCP) monthly analysis (new version 2.3) and a review of 2017 global precipitation. Atmosphere, 9, 138, https://doi.org/10.3390/atmos9040138.
Bailey, H., A. Hubbard, E. S. Klein, K.-R. Mustonen, P. D. Akers, H. Marttila, and J. M. Welker, 2021: Arctic sea-ice loss fuels extreme European snowfall. Nat. Geosci., 14, 283–288, https://doi.org/10.1038/s41561-021-00719-y.
Barnes, E. A., and J. A. Screen, 2015: The impact of Arctic warming on the midlatitude jet-stream: Can it? Has it? Will it? Wiley Interdiscip. Rev.: Climate Change, 6, 277–286, https://doi.org/10.1002/wcc.337.
Bintanja, R., and F. M. Selten, 2014: Future increases in Arctic precipitation linked to local evaporation and sea-ice retreat. Nature, 509, 479–482, https://doi.org/10.1038/nature13259.
Blackport, R., and P. J. Kushner, 2017: Isolating the atmospheric circulation response to Arctic sea ice loss in the coupled climate system. J. Climate, 30, 2163–2185, https://doi.org/10.1175/JCLI-D-16-0257.1.
Blackport, R., J. A. Screen, K. van der Wiel, and R. Bintanja, 2019: Minimal influence of reduced Arctic sea ice on coincident cold winters in mid-latitudes. Nat. Climate Change, 9, 697–704, https://doi.org/10.1038/s41558-019-0551-4.
Blackport, R., J. C. Fyfe, and J. A. Screen, 2022: Arctic change reduces risk of cold extremes. Science, 375, 729–729, https://doi.org/10.1126/science.abn2414.
Booth, J. F., C. M. Naud, and J. Willison, 2018: Evaluation of extratropical cyclone precipitation in the North Atlantic basin: An analysis of ERA-Interim, WRF, and two CMIP5 models. J. Climate, 31, 2345–2360, https://doi.org/10.1175/JCLI-D-17-0308.1.
Cassano, E. N., J. J. Cassano, M. E. Higgins, and M. C. Serreze, 2014: Atmospheric impacts of an Arctic sea ice minimum as seen in the Community Atmosphere Model. Int. J. Climatol., 34, 766–779, https://doi.org/10.1002/joc.3723.
Chang, E. K.-M., A. M.-W. Yau, and R. Zhang, 2022: Finding storm track activity metrics that are highly correlated with weather impacts. Part II: Estimating precipitation change associated with projected storm track change over Europe. J. Climate, 35, 2423–2440, https://doi.org/10.1175/JCLI-D-21-0259.1.
Christidis, N., and P. A. Stott, 2022: Human influence on seasonal precipitation in Europe. J. Climate, 35, 5215–5231, https://doi.org/10.1175/JCLI-D-21-0637.1.
Cohen, J., and Coauthors, 2020: Divergent consensuses on Arctic amplification influence on midlatitude severe winter weather. Nat. Climate Change, 10, 20–29, https://doi.org/10.1038/s41558-019-0662-y.
Cohen, J., L. Agel, M. Barlow, C. I. Garfinkel, and I. White, 2021: Linking Arctic variability and change with extreme winter weather in the United States. Science, 373, 1116–1121, https://doi.org/10.1126/science.abi9167.
Cvijanovic, I., B. D. Santer, C. Bonfils, D. D. Lucas, J. C. H. Chiang, and S. Zimmerman, 2017: Future loss of Arctic sea-ice cover could drive a substantial decrease in California’s rainfall. Nat. Commun., 8, 1947, https://doi.org/10.1038/s41467-017-01907-4.
Dai, A., D. Luo, M. Song, and J. Liu, 2019: Arctic amplification is caused by sea-ice loss under increasing CO2. Nat. Commun., 10, 121, https://doi.org/10.1038/s41467-018-07954-9.
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. Sun, R. A. Tomas, and J. Screen, 2016: 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.
Eady, E. T., 1949: Long waves and cyclone waves. Tellus, 1, 33–52, https://doi.org/10.3402/tellusa.v1i3.8507.
England, M. R., I. Eisenman, N. J. Lutsko, and T. J. W. Wagner, 2021: The recent emergence of Arctic amplification. Geophys. Res. Lett., 48, e2021GL094086, https://doi.org/10.1029/2021GL094086.
Eyring, V., S. Bony, G. A. Meehl, C. A. Senior, B. Stevens, R. J. Stouffer, and K. E. Taylor, 2016: Overview of the Coupled Model Intercomparison Project phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev., 9, 1937–1958, https://doi.org/10.5194/gmd-9-1937-2016.
Graff, L. S., and J. H. LaCasce, 2012: Changes in the extratropical storm tracks in response to changes in SST in an AGCM. J. Climate, 25, 1854–1870, https://doi.org/10.1175/JCLI-D-11-00174.1.
Hardiman, S. C., N. J. Dunstone, A. A. Scaife, D. M. Smith, R. Comer, Y. Nie, and H.-L. Ren, 2022: Missing eddy feedback may explain weak signal-to-noise ratios in climate predictions. npj Climate Atmos. Sci., 5, 57, https://doi.org/10.1038/s41612-022-00280-4.
Hay, S., and Coauthors, 2022: Separating the influences of low-latitude warming and sea ice loss on Northern Hemisphere climate change. J. Climate, 35, 2327–2349, https://doi.org/10.1175/JCLI-D-21-0180.1.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hodges, K. I., 1994: A general method for tracking analysis and its application to meteorological data. Mon. Wea. Rev., 122, 2573–2586, https://doi.org/10.1175/1520-0493(1994)122<2573:AGMFTA>2.0.CO;2.
Hodges, K. I., 1999: Adaptive constraints for feature tracking. Mon. Wea. Rev., 127, 1362–1373, https://doi.org/10.1175/1520-0493(1999)127<1362:ACFFT>2.0.CO;2.
Hoskins, B. J., and K. I. Hodges, 2002: New perspectives on the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 59, 1041–1061, https://doi.org/10.1175/1520-0469(2002)059<1041:NPOTNH>2.0.CO;2.
Kim, B.-M., S.-W. Son, S.-K. Min, J.-H. Jeong, S.-J. Kim, X. Zhang, T. Shim, and J.-H. Yoon, 2014: Weakening of the stratospheric polar vortex by Arctic sea-ice loss. Nat. Commun., 5, 4646, https://doi.org/10.1038/ncomms5646.
Kodama, C., B. Stevens, T. Mauritsen, T. Seiki, and M. Satoh, 2019: A new perspective for future precipitation change from intense extratropical cyclones. Geophys. Res. Lett., 46, 12 435–12 444, https://doi.org/10.1029/2019GL084001.
Landrum, L., and M. M. Holland, 2020: Extremes become routine in an emerging new Arctic. Nat. Climate Change, 10, 1108–1115, https://doi.org/10.1038/s41558-020-0892-z.
Lehmann, J., and D. Coumou, 2015: The influence of mid-latitude storm tracks on hot, cold, dry and wet extremes. Sci. Rep., 5, 17491, https://doi.org/10.1038/srep17491.
Lehmann, J., D. Coumou, K. Frieler, A. V. Eliseev, and A. Levermann, 2014: Future changes in extratropical storm tracks and baroclinicity under climate change. Environ. Res. Lett., 9, 084002, https://doi.org/10.1088/1748-9326/9/8/084002.
Levine, X. J., I. Cvijanovic, P. Ortega, M. G. Donat, and E. Tourigny, 2021: Atmospheric feedback explains disparate climate response to regional Arctic sea-ice loss. npj Climate Atmos. Sci., 4, 28, https://doi.org/10.1038/s41612-021-00183-w.
Liang, Y.-C., and Coauthors, 2020: Quantification of the Arctic sea ice-driven atmospheric circulation variability in coordinated large ensemble simulations. Geophys. Res. Lett., 47, e2019GL085397, https://doi.org/10.1029/2019GL085397.
Liu, J., D. Wu, X. Xu, M. Ji, Q. Chen, and X. Wang, 2021: Projection of extreme precipitation induced by Arctic amplification over the Northern Hemisphere. Environ. Res. Lett., 16, 074012, https://doi.org/10.1088/1748-9326/ac0acc.
Liu, J., M. Song, Z. Zhu, R. M. Horton, Y. Hu, and S.-P. Xie, 2022: Arctic sea-ice loss is projected to lead to more frequent strong El Niño events. Nat. Commun., 13, 4952, https://doi.org/10.1038/s41467-022-32705-2.
McCrystall, M. R., J. Stroeve, M. Serreze, B. C. Forbes, and J. A. Screen, 2021: New climate models reveal faster and larger increases in Arctic precipitation than previously projected. Nat. Commun., 12, 6765, https://doi.org/10.1038/s41467-021-27031-y.
McCusker, K. E., P. J. Kushner, J. C. Fyfe, M. Sigmond, V. V. Kharin, and C. M. Bitz, 2017: Remarkable separability of circulation response to Arctic sea ice loss and greenhouse gas forcing. Geophys. Res. Lett., 44, 7955–7964, https://doi.org/10.1002/2017GL074327.
McKenna, C. M., T. J. Bracegirdle, E. F. Shuckburgh, P. H. Haynes, and M. M. Joshi, 2018: Arctic sea ice loss in different regions leads to contrasting Northern Hemisphere impacts. Geophys. Res. Lett., 45, 945–954, https://doi.org/10.1002/2017GL076433.
Mioduszewski, J., S. Vavrus, and M. Wang, 2018: Diminishing Arctic sea ice promotes stronger surface winds. J. Climate, 31, 8101–8119, https://doi.org/10.1175/JCLI-D-18-0109.1.
Monerie, P.-A., T. Oudar, and E. Sanchez-Gomez, 2019: Respective impacts of Arctic sea ice decline and increasing greenhouse gases concentration on Sahel precipitation. Climate Dyn., 52, 5947–5964, https://doi.org/10.1007/s00382-018-4488-5.
Mori, M., M. Watanabe, H. Shiogama, J. Inoue, and M. Kimoto, 2014: Robust Arctic sea-ice influence on the frequent Eurasian cold winters in past decades. Nat. Geosci., 7, 869–873, https://doi.org/10.1038/ngeo2277.
Notz, D., and J. Stroeve, 2016: Observed Arctic sea-ice loss directly follows anthropogenic CO2 emission. Science, 354, 747–750, https://doi.org/10.1126/science.aag2345.
Notz, D., and SIMIP Community, 2020: Arctic sea ice in CMIP6. Geophys. Res. Lett., 47, e2019GL086749, https://doi.org/10.1029/2019GL086749.
Oudar, T., E. Sanchez-Gomez, F. Chauvin, J. Cattiaux, L. Terray, and C. Cassou, 2017: Respective roles of direct GHG radiative forcing and induced Arctic sea ice loss on the Northern Hemisphere atmospheric circulation. Climate Dyn., 49, 3693–3713, https://doi.org/10.1007/s00382-017-3541-0.
Peings, Y., Z. M. Labe, and G. Magnusdottir, 2021: Are 100 ensemble members enough to capture the remote atmospheric response to +2°C Arctic sea ice loss? J. Climate, 34, 3751–3769, https://doi.org/10.1175/JCLI-D-20-0613.1.
Priestley, M. D. K., and J. L. Catto, 2022: Future changes in the extratropical storm tracks and cyclone intensity, wind speed, and structure. Wea. Climate Dyn., 3, 337–360, https://doi.org/10.5194/wcd-3-337-2022.
Rantanen, M., A. Y. Karpechko, A. Lipponen, K. Nordling, O. Hyvärinen, K. Ruosteenoja, T. Vihma, and A. Laaksonen, 2022: The Arctic has warmed nearly four times faster than the globe since 1979. Commun. Earth Environ., 3, 168, https://doi.org/10.1038/s43247-022-00498-3.
Rienecker, M. M., and Coauthors, 2011: MERRA: NASA’s Modern-Era Retrospective Analysis for Research and Applications. J. Climate, 24, 3624–3648, https://doi.org/10.1175/JCLI-D-11-00015.1.
Rodgers, K. B., and Coauthors, 2021: Ubiquity of human-induced changes in climate variability. Earth Syst. Dyn., 12, 1393–1411, https://doi.org/10.5194/esd-12-1393-2021.
Saha, S., and Coauthors, 2010: The NCEP Climate Forecast System Reanalysis. Bull. Amer. Meteor. Soc., 91, 1015–1057, https://doi.org/10.1175/2010BAMS3001.1.
Screen, J. A., and I. Simmonds, 2010: The central role of diminishing sea ice in recent Arctic temperature amplification. Nature, 464, 1334–1337, https://doi.org/10.1038/nature09051.
Screen, J. A., C. Deser, I. Simmonds, and R. Tomas, 2014: Atmospheric impacts of Arctic sea-ice loss, 1979–2009: Separating forced change from atmospheric internal variability. Climate Dyn., 43, 333–344, https://doi.org/10.1007/s00382-013-1830-9.
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.
Screen, J. A., and Coauthors, 2018: Consistency and discrepancy in the atmospheric response to Arctic sea-ice loss across climate models. Nat. Geosci., 11, 155–163, https://doi.org/10.1038/s41561-018-0059-y.
Screen, J. A., R. Eade, D. M. Smith, S. Thomson, and H. Yu, 2022: Net equatorward shift of the jet streams when the contribution from sea-ice loss is constrained by observed eddy feedback. Geophys. Res. Lett., 49, e2022GL100523, https://doi.org/10.1029/2022GL100523.
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.
Shen, Z., A. Duan, D. Li, and J. Li, 2021: Assessment and ranking of climate models in Arctic Sea ice cover simulation: From CMIP5 to CMIP6. J. Climate, 34, 3609–3627, https://doi.org/10.1175/JCLI-D-20-0294.1.
Shu, Q., Q. Wang, Z. Song, F. Qiao, J. Zhao, M. Chu, and X. Li, 2020: Assessment of sea ice extent in CMIP6 with comparison to observations and CMIP5. Geophys. Res. Lett., 47, e2020GL087965, https://doi.org/10.1029/2020GL087965.
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.
Smith, D. M., and Coauthors, 2019: The Polar Amplification Model Intercomparison Project (PAMIP) contribution to CMIP6: Investigating the causes and consequences of polar amplification. Geosci. Model Dev., 12, 1139–1164, https://doi.org/10.5194/gmd-12-1139-2019.
Smith, D. M., and Coauthors, 2022: Robust but weak winter atmospheric circulation response to future Arctic sea ice loss. Nat. Commun., 13, 727, https://doi.org/10.1038/s41467-022-28283-y.
Stroeve, J., and D. Notz, 2018: Changing state of Arctic sea ice across all seasons. Environ. Res. Lett., 13, 103001, https://doi.org/10.1088/1748-9326/aade56.
Sun, L., C. Deser, and R. A. Tomas, 2015: Mechanisms of stratospheric and tropospheric circulation response to projected Arctic sea ice loss. J. Climate, 28, 7824–7845, https://doi.org/10.1175/JCLI-D-15-0169.1.
Tomas, R. A., C. Deser, and L. Sun, 2016: The role of ocean heat transport in the global climate response to projected Arctic sea ice loss. J. Climate, 29, 6841–6859, https://doi.org/10.1175/JCLI-D-15-0651.1.
Vihma, T., 2014: Effects of Arctic sea ice decline on weather and climate: A review. Surv. Geophys., 35, 1175–1214, https://doi.org/10.1007/s10712-014-9284-0.
Wang, M., and J. E. Overland, 2009: A sea ice free summer Arctic within 30 years? Geophys. Res. Lett., 36, L07502, https://doi.org/10.1029/2009GL037820.
Wu, B., K. Yang, and J. A. Francis, 2017: A cold event in Asia during January–February 2012 and its possible association with Arctic sea ice loss. J. Climate, 30, 7971–7990, https://doi.org/10.1175/JCLI-D-16-0115.1.
Wu, P., N. Christidis, and P. Stott, 2013: Anthropogenic impact on Earth’s hydrological cycle. Nat. Climate Change, 3, 807–810, https://doi.org/10.1038/nclimate1932.
Xu, M., W. Tian, J. Zhang, J. A. Screen, J. Huang, K. Qie, and T. Wang, 2021: Distinct tropospheric and stratospheric mechanisms linking historical Barents-Kara sea-ice loss and late winter Eurasian temperature variability. Geophys. Res. Lett., 48, e2021GL095262, https://doi.org/10.1029/2021GL095262.
Zhang, R., and J. A. Screen, 2021: Diverse Eurasian winter temperature responses to Barents-Kara sea ice anomalies of different magnitudes and seasonality. Geophys. Res. Lett., 48, e2021GL092726, https://doi.org/10.1029/2021GL092726.
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.
