A review of the interaction between the cryosphere and atmosphere

YongJian Ding,JianPing Yang,ShengXia Wang,YaPing Chang

1. State Key Laboratory of Cryospheric Science, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences,Lanzhou,Gansu 730000,China

2.China-Pakistan Joint Research Center on Earth Science,CAS-HEC,Islamabad,Pakistan

3.University of Chinese Academy of Sciences,Beijing 100049,China

ABSTRACT The interaction between the cryosphere and atmosphere is an essential and extremely sensitive mutual action process on the earth. Due to global warming and the cryospheric melting, more and more attention has been paid to the interaction process between the cryosphere and atmosphere, especially the feedback of the cryosphere change to the atmosphere. A comprehensive review of the studies on the interaction between the cryosphere and atmosphere is conducted from two as‐pects: (1) effects of climate change on the cryosphere or responses of the cryosphere to climate change; and (2) feedback of the cryosphere change to the climate.The response of the cryosphere to climate change is lagging. Such a lagging and cumulative effect of temperature rise within the cryosphere have resulted in a rapid change in the cryosphere in the 21st century,and its impacts have become more significant.The feedback from cryosphere change on the climate are omnifari‐ous.Among them, the effects of sea ice loss and snow cover change, especially the Arctic sea ice loss and the Northern Hemisphere snow cover change, are the most prominent. The Arctic amplification (AA) associated with sea ice feedback is disturbing,and the feedback generated by the effect of temperature rise on snow properties in the Northern Hemisphere is also of great concern.There are growing evidence of the impact of the Arctic cryosphere melting on mid-latitude weath‐er and climate.Weakened storm troughs,steered jet stream and amplified planetary waves associated with energy propaga‐tion become the key to explaining the links between Arctic cryosphere change and atmospheric circulation.There is still a great deal of uncertainty about how cryosphere change affects the weather and climate through different atmospheric circu‐lation processes at different spatial and temporal scales due to observation and simulation problems.

Keywords:climate warming;cryosphere;atmosphere;response;feedback

1 Introduction

The cryosphere is composed of glaciers, ice caps,snow cover,frozen soil, river and lake ice, sea ice, ice shelf, iceberg, and frozen water in the atmosphere.The term "Cryosphere" was formally proposed as the fifth earth sphere alongside the atmosphere, hydro‐sphere, biosphere and lithosphere by the World Mete‐orological Organization(WMO)at the Stockholm Hu‐man Environment Conference in 1972. It is a crucial element of the earth's environmental systems. Of the environmental systems,the cryosphere is the most sig‐nificant and indicative, and the fastest responding to global change. It is also the most direct and sensitive to the climate system (Ding and Xiao, 2013). Thus it is acknowledged as one of the core links and key fac‐tors of multi-sphere interactions in the climate system(IPCC, 2007). Meanwhile, as a component of the five spheres of the climate system, the interactions be‐tween the cryosphere, atmosphere, hydrosphere, bio‐sphere and lithosphere(mainly the surface part)are al‐so the focus of cryosphere science(Qinet al.,2018).

What is the interaction between the cryosphere and other spheres? It is defined as the crossing parts in which the cryosphere plays a major role in the inter‐linkages and interactions between the cryosphere and other spheres (Figure 1). It is evident that the study of the interactions between the cryosphere and other earth spheres is interdisciplinary. The focus of the in‐terdisciplinary study is on the cryosphere, and the in‐vestigations involve the closely correlated crossing parts, excluding the fields other than the inter-sphere cross and those areas which are not directly related to the cryosphere itself.

Figure 1 Interactions between the cryosphere and other earth spheres

There are many literatures and textbooks on the sciences of the cryosphere itself. However, the inter‐actions with other earth spheres and the identifica‐tion and quantification of the cryosphere's roles dur‐ing the interacting processes are the hotspot of the re‐cent cryosphere research and attract widespread at‐tention. The cryosphere, by interacting with other earth spheres, can affect global energy balance, water balance, carbon-nitrogen balance, water-salt balance,thermal-salt balance,etc.. It also can affect weather and climate,water cycle and water resources,terrestri‐al and marine ecosystems, surface environment and disasters through changes in energy and water, heat and salt, and biogeochemistry, thereby affecting sus‐tainable human development (Dinget al., 2012;French and Slaymaker, 2012; Levermannet al.,2012; Steffenet al., 2012; Ding and Xiao, 2013;Vaughanet al., 2013; Field, 2014; AMAP, 2017;Richter-Mengeet al., 2017; Qinet al., 2018; Boyet al.,2019)(Figure 1).

The study of the interaction between the cryo‐sphere and the atmosphere mainly includes two as‐pects: the effect of climate change on the cryosphere and the feedback of cryosphere change on weather and climate. The impact of climate change on the cryosphere, that is, the response of the cryosphere to climate change, is manifested in various spatial and temporal scales and important change links of the ef‐fects of climate change on cryosphere formation and evolution. The latter is also the traditional research content of the cryosphere. Changes in the cryo‐sphere, in turn, can affect global climate change,which is the feedback of the cryosphere on the cli‐mate. There are many feedback from the cryosphere to the climate, such as ice and snow albedo feed‐back, radiative feedback, greenhouse gas emission increase from permafrost thawing,etc.. Callaghanet al. (2011) have synthesized the major feedback on the climate mediated by cryospheric elements, includ‐ing sea ice, Greenland ice sheet, snow, permafrost,glaciers and ice caps,and lake and river ice.The feed‐back of the cryosphere change on the climate has at‐tracted widespread attention and has become a re‐search hot topic in atmospheric science and cryo‐sphere science in the last decade. For the former, this paper only reviews some new features of the cryo‐spheric response to climate change. For the latter, the focus and hot issues are reviewed from the aspects of cryospheric feedback on the climate and the relation‐ship between high-latitude cryosphere change and mid-latitude weather and climate.

2 Response of the cryosphere to climate change

Mutual feedback and response between the cryo‐sphere and climate are closely related to different cryospheric elements and size. For instance, on the global scale, mountain glaciers respond significantly to the climate on a decadal scale (Ding, 1996). On a centennial scale, the retreat of mountain glaciers has provided clear categorical evidence of climate change (Roeet al., 2016). However, the response of glaciers to the climate is nonlinear, and the response process is more complicated. For example, global warming has lasted for hundreds of years and accel‐erated in the 20th century.Still,the loss of global gla‐cier mass in the second half of the 20th century was only slightly higher than that in the first half (Lecler‐cqet al.,2011;Marzeionet al.,2012).Moreover,be‐cause glacier response times are typically decades or longer, the present-day glacier retreat is a mixed re‐sponse to past and current natural climate variability and current anthropogenic forcing (Marzeionet al.,2014a). Simulations of the global glacier mass bal‐ance changes indicate that about 75% of the global glacier mass loss from 1851 to 2010 is caused by nat‐ural climate variability, and only about 25% was at‐tributable to human emission-induced climate warm‐ing. However, the contribution of warming caused by greenhouse gases has increased significantly since the late 20th century. The anthropogenic frac‐tion of global glacier mass loss has grown to nearly 70% from 1991 to 2010 (Marzeionet al., 2014a,b).The Antarctic is slow to respond to climate change because of its colossal size, but its impact is enor‐mous. A study of the relationship between Antarctic ice sheet and climate change shows that fluctuations in Antarctic ice sheet discharge caused by relatively small changes in subsurface ocean temperature can amplify multi-centennial climate variability regional‐ly and globally, suggesting that a dynamic Antarctic ice sheet may have driven climate fluctuations dur‐ing the Holocene (Bakkeret al., 2016). As for the contribution to sea-level rise (SLR), the Antarctic has the potential to contribute more than 1 m to SLR by 2100, outpacing Arctic glaciers and ice caps (Boxet al., 2017). Meanwhile, a range of weather and cli‐mate,hydrological and other effects to SLR are poor‐ly understood.

The above-mentioned study results indicate that cryospheric response to climate change is lagging.The lag time is not only related to the intensity and rate of global climate warming but also to the "cumu‐lative temperature rise" effect of continuous warming within the cryosphere. The effects are reflected in the magnitude and spatial-temporal scale of cryosphere change through thermal and dynamic processes with‐in the cryosphere(Dinget al.,2019).

Studies on the relationship between Eurasia snow cover and air temperature and precipitation show that the most substantial contribution of air temperature to snow accumulation is mainly found in those months with moderate mean air temperature. The most deci‐sive contribution of precipitation is not collocated with the climatological maxima in the same. And a spatial dipole characterizes the leading mode of March snow cover extension variability.These charac‐teristics are related to anomalies in atmospheric water vapor divergence,storm activity and the associated at‐mospheric circulation (Ye, 2019). At present, re‐sponse study of snow cover to the climate is mainly concerned with changes in snow cover range, snow depth, and so on. In contrast, the change of snow physical properties, such as snow particle size, water content, surface and internal impurities content, and its climatic and hydrological effect are not capturing sufficient attention.

Permafrost response to climate change is compli‐cated. Under climate warming, for warm permafrost(above −1.0°C),the heat entering the ground is main‐ly consumed to thaw permafrost because the tempera‐ture of permafrost is close to the ice melting point(Figure 2b). In this case, the lowering of the perma‐frost table, the increased active layer thickness, and weak heat conduction to the permafrost result in a very shallow depth of zero annual amplitude (DZAA)of the ground temperature. For cold permafrost, due to the large cold storage in it, the heat transferred into the ground warms the permafrost, and a large amount of heat will diffuse into the permafrost(Figure 2a).As a result, DZAA moves deeper, leading to slower per‐mafrost degradation and smaller change in the active layer. Once the permafrost becomes warm, DZAA will quickly move to a shallower depth, accelerating the ground thaw. Dinget al.(2019) analyzed the rela‐tionships between the rate of active layer change(RALC) and the mean annual ground temperature(MAGT) at a depth of 10 to 20 m (Figure 2c), and found that there was no obvious correlation between cold permafrost, warm permafrost and active layer change. With the increase of permafrost temperature,RALC displayed an obvious upward trend (Figure 2d).When MAGT was around −2 °C,RALC changed slowly as MAGT increased within the low-tempera‐ture ranges. RALC rose rapidly with an increase in MAGT within the high-temperature ranges, such as above −2 °C. Such results reflect a correlation be‐tween permafrost temperature and active layer change across the world. In other words, whether the climate warming will lead to a significant permafrost retreat mainly depends on its thermal state and the extent of temperature increase. The transition from cold perma‐frost to warm one is a warning sign for a complete retreat(Dinget al.,2019).

Figure 2 Response pattern of permafrost to climate warming.(a)cold permafrost,(b)warm permafrost,(c)RALC and MAGT data at 103 observational stations across the world;(d)classification of the data of MAGT according to the levels of−0.5 to −1.0,−1.0 to −1.5,−1.5 to −2.0,−2.0 to −2.5,−2.5 to −3.0,−3.0 to −4.0,−4.0 to −5.0,−5.0 to −6.0,−6.0 to−7.0,<−7.0°C,and averaging the MAGT and RALC.The fluctuation range of RALC at the MAGT 0 to−0.5°C was the largest,with the mean value of 12.98 cm/a（Ding et al.,2019）

3 Feedback from cryosphere change to the climate

3.1 Feedback of cryosphere change

The feedback of cryosphere change on the climate is mainly reflected in the variations of the global or re‐gional energy and water balance, which in turn affects the global energy and water distribution and eventually leads to the global climate change(Figure 3).These ef‐fects range from the land surface energy balance change due to land snow cover and permafrost to the impacts of ice sheet and sea ice on global reflectance and ocean thermohaline circulation. The spatial range includes global scale (ocean and land), hemispheric scale (Northern Hemisphere snow cover change), re‐gional scale (Tibetan Plateau cryosphere, polar cryo‐sphere),and so on.The time scale varies from season‐al, interannual, interdecadal to centennial, millennial or even ten-thousand years (Ding and Xiao, 2013;Vaughanet al., 2013;AMAP, 2017; Richter-Mengeet al., 2017; Boyet al., 2019). For different cryospheric elements such as glaciers, snow cover, sea ice,etc.,scales of their feedback on the climate are different.Glacier change affects local energy and water circula‐tion through surface albedo and alters mountainous en‐ergy and water internal circulation (Dinget al., 2017;Ding and Zhang, 2018; Chenet al., 2019). Ablation and reduction of snow cover can increase radiation ab‐sorption at various spatial scales and enhance albedo's feedback process.Melting of ice sheet not only causes sea level to rise but also affects thermohaline circula‐tion, and in turn, the global climate through the en‐hanced freshwater effect.The ice sheet change mainly affects climate change at longer time scales (Bakkeret al., 2016). The reduction in sea ice increases ocean radiation absorption, modifies the albedo feedback of ice, escalates atmospheric heat and water vapor ener‐gy, and influences atmospheric cloudiness (Callaghanet al.,2011).Permafrost change affects energy and wa‐ter exchanges between land and atmosphere through latent heat, changes the climate through greenhouse gas emissions, and alters the surface heat by changing surface water moisture (wetland drying, thermal melt‐ing karst, vegetation coverage change,etc.). Change in lake ice impacts local atmospheric processes at dai‐ly and monthly scales, and even throughout the freez‐ing period (Boyet al., 2019). Change in the cloud cover influences the radiation input and expenditure of the cryosphere surface, which in turn, influences the cloud cover forming a cloud radiation-cryosphere feedback process(Callaghanet al.,2011).

3.2 Arctic amplification

The albedo feedback in the cryosphere change is prominent in the Arctic region, which is commonly known as"Arctic amplification".As the climate warms,Arctic amplification is manifested as a decrease in the range of ice and snow and the albedo, and an increase in surface heat absorption, thus further accelerating cli‐mate warming. The interaction relationships between Arctic sea ice and snow cover and climate change are more closely (Serreze and Barry, 2011). The decreases in polar snow cover and sea ice lead to an increase in absorbed solar radiation in summer. The lesser longwave radiation loss reduces the temperature difference between the Arctic region and the middle and low lati‐tudes, which further escalates climate warming (Pithan and Mauritsen, 2014; Boyet al., 2019). Relatively speaking,since the Arctic cryosphere's surface tempera‐ture is closer to the melting point, its albedo feedback is particularly sensitive, and its climate feedback effect is more prominent than that of the Antarctic.In the Arc‐tic, however, the importance of different feedback to climate change, such as the effects of sea ice and snow cover and their seasonal amplifications,remains contro‐versial (Hudson, 2011; Perovichet al., 2012).At pres‐ent, many factors have been found to have an impor‐tant impact on Arctic amplification, including sea ice loss, changes in atmospheric and marine thermal ener‐gy, changes in cloudiness and water vapor content, sea ice output, and atmospheric circulation channels that are closely associated with Arctic and mid-latitude weather,etc.(Cohenet al.,2014).

The melting of sea ice leads to the expansion of open sea areas and the formation of more breaking waves and foams,increased seafoam aerosols,biogenic sulfide dimethyl sulfur (DMS) and secondary organic aerosol precursors (Figure 3) (Norriset al., 2011).These outcomes further affect atmospheric aerosol concentration and cloud amount(Boyet al.,2019).

3.3 The albedo feedback of snow cover

Because of its high albedo,snow cover has a great influence on surface radiation.And snow particle size is the main factor that determines the snow albedo(Domineet al., 2006). Air temperature has an impor‐tant effect on snow particle size. The higher the tem‐perature, the larger the snow particle size will become by metamorphism. Larger snow crystals increase the optical channel to absorb more energy, accelerating the snow melting,reducing the surface albedo,and en‐hancing the land surface radiation absorption. As a result, throughout the Northern Hemisphere, warm‐ing-induced changes in snow cover form an increas‐ing snow-albedo positive feedback process (Boyet al., 2019). Though remote sensing analysis showed that the snow cover range in the Northern Hemisphere remained largely unchanged from 2000 to 2012, the albedo changed ±0.2 (Atlaskinaet al., 2015), and the variation is mainly attributed to the temperature change. Laboratory study results have revealed that the snow surface albedo decreases when the air tem‐perature reaches −5 °C and above (Aokiet al., 2003).However,recent studies in the Arctic region show that the snow albedo-feedback between the atmosphere and cryosphere seems unusual, as evidenced by the positive feedback of snow-albedo in the Arctic region at very low temperatures and with complete snow cov‐er (Boyet al., 2019). This result indicates high sensi‐tivity of the Arctic region to global warming and more rapid climate change to come. Moreover, aerosol par‐ticles and impurities such as black carbon (BC), or‐ganic carbon(OC),dust and microorganisms in the at‐mosphere fall on the snow surface (Figure 3), affect‐ing the snow albedo and the rate of snow melting and forming another short-term feedback process (Mein‐anderet al., 2013; Groot Zwaaftinket al., 2016; Boyet al.,2019;Dagsson-Waldhauserovaet al.,2019).

Snow cover has multiple driving and feedback ef‐fects. For example, climate warming increases soil moisture availability, alters atmospheric circulation, af‐fects vegetation, increases melting in winter and freez‐ing rain-snow events,etc. These driving and feedback effects between snow cover and the climate can occur at different time and spatial scales and vary significant‐ly over the seasons with snow cover range, accumula‐tion time and snow physical properties. It is generally accepted that snow cover anomalies over Eurasia in October acted to reinforce the Siberian High and the negative mode of the Arctic Oscillation (AO) through a vertically propagating Rossby wave train (Cohen and Entekhabi 1999; Fletcheret al., 2009; Allen and Zender, 2011; Cohenet al., 2012; Liet al., 2017). A recent study by Orsoliniet al.(2016) indicated that snow depth anomalies, rather than snow cover exten‐sion, are responsible for initiating and maintaining an ongoing anomalously negative AO phase. This study found that the Tibetan Plateau snow cover (TPSC) is closely linked to the interannual variations of summer heat waves over Eurasia, which helps to explain more than 30%of the total variances of heat wave variability in southern Europe and northeastern Asia (Wuet al.,2016). Meanwhile, the TPSC has more impact on sub‐sequent weather and climate systems, including the East and South Asian Summer Monsoons and their sub‐sequent precipitation regimes (Youet al., 2020). The sub-seasonal variability of TPSC is also closely related to the subsequent East Asian atmospheric circulation at medium-range time scales (approximately 3 −8 days later) during wintertime. TPSC acts as an elevated cooling source in the middle troposphere during winter‐time and rapidly modulates the land surface thermal conditions over the Tibetan Plateau(Liet al.,2018).

The studies on the relationship between Eurasian spring snow decrement and East Asian summer pre‐cipitation show(Zhanget al.,2017;Zhanget al.,2019)that a west−east dipole pattern in Eurasian spring snow decrement anomalies, with a negative center located in the region between eastern Europe and the West Si‐beria Plain and a positive center located around Bai‐kal Lake, is closely associated with East Asian sum‐mer precipitation via an anomalous midlatitude Eur‐asian wave train. Reduced spring snow decrement over Eastern Europe and the West Siberia Plain corre‐sponds to anomalously dry local soil conditions from spring to the following summer, thereby increasing surface heat flux and near-surface temperatures. Fur‐ther studies (Ye, 2019) indicate that the dipole pattern of November Eurasian snow cover seems to be forced by the autumn Arctic sea ice concentration (SIC) over the Barents and Kara Seas (B/K), and its feedback to the atmospheric circulation is important. Therefore,the impacts of autumn B/K Sea SIC on the autumn and wintertime atmospheric circulation and thus the March snow water equivalent variability may be mod‐ulated by both constructive and destructive interfer‐ences of autumn Eurasian snow cover. Studies on snow cover in Southern Asia indicate (Zhanget al.,2019) that the relationship between Eurasian spring snow cover and Indian summer monsoon rainfall is weakening, while inverse relationship between the In‐dian summer monsoon rainfall and central Eurasian spring snow cover has disappeared since 1990.

In conclusion,the high albedo of ice and snow sur‐face, the coexistence of ice, water and water vapor,and the large amount of latent heat in the process of phase transition are the key links of the interaction be‐tween the cryosphere and climate.These very process‐es result in the strong influence of the cryosphere on the surface energy balance. The global scale cryo‐sphere, whose existence or not is directly correlated with the temperature difference, and then affects the wind strength and the ocean thermohaline circulation(French and Slaymaker, 2012). With continuous cli‐mate warming, the impacts of the cryosphere on the climate will become more and more prominent, but the net effect is still difficult to assess, since the spa‐tial and temporal scales of different cryosphere ele‐ments acting on the climate system vary greatly. The existing general circulation models(GCMs)do not in‐clude all major feedback from the cryosphere to the climate system, such as the feedback between perma‐frost and vegetation and the feedback of the changes in the physical properties of snow cover to the atmo‐sphere (Brownet al., 2017). Even if the cryosphere feedback is considered in some GCMs, its parameter‐ization accuracy is not high. There is a lack of full coupling between the surface dynamics of the cryo‐sphere and the atmosphere,which is the main gap that currently restricts the reduction of uncertainty in GCMs(Callaghanet al.,2011).

4 Impacts of the Arctic cryosphere change on mid- and low-latitude weather and climate

4.1 Mechanism of the Arctic cryosphere change affecting medium-latitude atmospheric circulation

More and more studies have shown that Arctic cryosphere changes can affect mid-latitude weather and even the Southeast Asian monsoon.Arctic ampli‐fication (AA) is mainly affected by the feedback from cryosphere change altering surface energy. At the same time, the effect of AA on midlatitude weather and climate is a specific manifestation of the cryo‐sphere influencing the whole globe by changing atmo‐spheric circulation. The global climate model simula‐tions have shown that AA is an important driving force of mid-latitude weather (AMAP, 2017). AA is related to many factors, such as ice and snow albedo,temperature gradients, Planck effect, CO2, cloud amount, atmospheric and ocean transports,etc.(Tay‐loret al., 2013; Pithan and Mauritsen, 2014). Howev‐er, ample evidence has shown that the main contribu‐tion of AA is an increase in earth's surface absorption energy resulted from the decrease in ice and snow al‐bedo. In particular, sea ice reduction has multiplied the heat absorbed by the oceans.Additionally, it is al‐so important to reconginze that the warming causes different vertical thermal structure differences be‐tween high and low latitudes, which is the result, not the cause,of changes in ice and snow albedo.Another important contribution of AA is the feedback of water vapor, which tends to go together with the direct radi‐ation effect caused by the increase in CO2concentra‐tion. The net effect of the cloud cover is relatively small, but the effect is greater in certain seasons of some years. The ocean heat transport has a negative feedback effect inhibiting Arctic warming (AMAP,2017).

The Arctic's influence mechanism on the weather and climate in the middle and low latitudes can be ex‐plained by synthesizing the aforementioned feedback(Overlandet al., 2015). The Arctic temperature rising rate is 2−3 times that of the middle latitudes.The large amplitude of warming can reduce the South −North temperature difference, which is the main driving force of the polar vortex or the polar jet stream. Thus the prevailing westerly jet stream is weakened (Over‐land and Wang, 2010; Cvijanovic and Caldeira, 2015;Francis and Vavrus, 2015). As a result, large-ampli‐tude planetary waves in the polar jet stream tend to oc‐cur more slowly, forming a more durable north-south channel and turning a southward average latitudinal jet. This wavier mode allows warm air to reach far‐ther away northward and cold air to invade southward to lower latitudes (Figure 4). When this weather sys‐tem is formed, it can be steadily sustained in one place, triggering extreme weather events (Screen and Simmonds, 2014). Saharan dust is transported to the Arctic during such an extreme event (Franciset al.,2018). Since 2007, the Arctic has experienced several unusually warm winters. The subsequent sea ice loss has led to a significant increase in the Siberian High,causing cold storms that affected most areas of China,South Korea and Japan (Petoukhov and Semenov,2010; Li and Wang, 2013; Kimet al., 2014; Kugetal., 2015; Semenov and Latif, 2015; Sunet al., 2016;Zhou,2017;Zhanget al.,2020).

Figure 4 Grammatic sketch link between Arctic amplification and weather and climate in midlatitude.(a)Stable state,(b)Unstable state

Further, a series of possible links between the above-mentioned Arctic melting and mid-latitude weather can be attributed to three atmospheric assump‐tions that are closely related to the effects of sea ice and snow cover changes: weakened storm troughs, steered jet stream and amplified planetary waves associated with their energy propagation (Cohenet al., 2014;Coumouet al., 2018). A simulation test for polar jet stream shows that when global climate warms, en‐hanced northward heat transport provides a major contribution to decrease the northward temperature gradient in the polar troposphere in cold seasons,causing more oscillation of the planetary waves (Me‐leshkoet al.,2016).Although AA significantly enhanc‐es near-surface air temperature in the polar region, it is not large enough to invoke an increased oscillation of the planetary waves. An intermittent stratospheric pathway whereby reduced sea ice in the Barents-Kara Seas enhances upward wave activity and wave-break‐ing in the stratosphere,leading to weakened polar vor‐tex and a transition of the North Atlantic Oscillation(NAO)to its negative phase(Siewet al.,2019).

The study by Kretschmeret al.(2016) suggests that the effect of the Arctic sea ice anomaly on the mid-latitude circulation is more important than that of snow cover. However, Furtadoet al.(2016) argued that autumnal sea ice and snow cover anomalies inter‐act differently with the phase and amplitude of the Northern Annular Mode. Both are required to produce skillful forecasts of mid-latitude winter temperature anomalies. It is difficult to judge which one has a stronger effect on the climate.Snow cover is distribut‐ed in large regions of the Northern Hemisphere. Its variation is seasonal and mainly related to the atmo‐sphere through the land surface energy feedback pro‐cess. In contrast, sea ice, including both one-year and multi-year, mainly occurs in high latitudes, with a smaller area than snow cover. It is directly related to the ocean and affects the weather and climate through ice-sea-atmosphere interaction. Therefore, the effect of the organic coupling of snow cover and sea ice on the spatial and temporal scales with ocean and atmo‐sphere is the key to understand the current climate feedback mechanism of cryosphere change. Climate system change is very complex, and often many pro‐cesses occur simultaneously.The upper tropical atmo‐sphere is also a warming amplification region, which,opposite to AA, helps to increase the north-south temperature difference between the tropics and the mid-latitudes (McCuskeret al., 2017; Peingset al.,2018). How the cold Arctic and hot tropics work to‐gether during climate warming and influence mid-lati‐tude weather and climate and extreme events through atmospheric circulation is still a mystery.

4.2 Links between Arctic sea ice loss and mid-latitude weather processes

The relationship between sea ice loss and extreme winter weather in Eurasia and summer continental at‐mospheric modes has received more and more atten‐tion in recent years (Francis and Vavrus, 2012; Li and Wang,2013;Guoet al.,2014;Moriet al.,2014;Tanget al., 2014; Petrieet al., 2015; McCuskeret al.,2016; Sunet al., 2016; Wadhams, 2016; Mannet al.,2017; Vavruset al., 2017; Yaoet al., 2017; Francis,2018;Kretschmeret al.,2018;Liet al.,2018;Vavrus,2018; Yeet al., 2018; Yorket al., 2018; Zhanget al.,2018; Blackport and Screen, 2019; Blackportet al.,2019; Boyet al., 2019; Daiet al., 2019; Luet al.,2019; Zhanget al., 2020). Concerning the relation‐ship between sea ice loss and cold winter in Eurasia at mid-latitudes in the wintertime, it has been found that sea ice loss in the Barents/Kara Seas helps to enhance the region's climatological ridge, further strengthening the Siberian High and transporting cold Arctic air to East Asia. Wave energy from a more meandering jet stream disrupts the stratospheric jet stream and rein‐forces the pattern into late winter(Li and Wang,2013;Sunet al., 2016; Yaoet al., 2017; Kretschmeret al.,2018; Liet al., 2018; Yeet al., 2018; Zhanget al.,2018; Luet al., 2019; Zhanget al., 2020). Moriet al.(2014) used a 100-member ensemble of simulations with an atmospheric general circulation model driven by observation-based sea ice concentration anomalies and showed that the probability of severe winters has more than doubled in central Eurasia as a result of sea ice loss in the Barents −Kara Seas. Reanalysis data and the simulations suggest that sea ice loss leads to more frequent Eurasian blocking situations, which se‐quentially favor cold-air advection to Eurasia and cause severe winters.Atmosphere-only global climate models, complemented by 50 ensembles of atmo‐spheric-ocean global climate model, as an internal variability and an external forcing change, are used to isolate the effects of sea ice reduction from climate change over the past 600 years. It has been found that the cold winter in Eurasia since the 1980s stems from the circulation modes generated in the Barents-Kara Seas and its nearby sea-ice-dependent (McCuskeret al.,2016).

After comparing the differences between observa‐tion and simulation and the underestimation error of sea ice climate effect in the past, it is found that ap‐proximately 44% of the central Eurasian cooling trend for 1995 −2014 is attributable to sea ice loss in the Barents−Kara Seas(Moriet al.,2019).But some stud‐ies have suggested that the role of AA is limited to a certain space-time range. Large AA occurs only from October to April and only over areas with significant sea ice loss (Daiet al., 2019). AA largely disappears when Arctic sea ice is fixed or melts (Daiet al.,2019).AA and sea ice loss increase precipitation and snow‐fall above approximately 60°N and reduce meridional temperature gradients above approximately 45°N in the lower−mid troposphere. However, minimal impact on the mean climate is seen below approximately 60°N,with weak reduction in zonal wind over 50°N−70°N and 150−700 hPa, mainly over the North Atlantic and northern central Asia(Dai and Song，2020).Recent stud‐ies suggested that severe midlatitude winters caused by the winter atmospheric circulation response to sea ice loss are primarily driven by sea ice loss in winter rather than in Autumn (Blackport and Screen, 2019;Blackportet al., 2019). They found that the correla‐tion between reduced sea ice and extreme winters across the mid-latitude occurs because both are simul‐taneously driven by the same, large-scale atmospheric circulation patterns. Crucially, it shows that reduced sea ice only has a minimal influence on whether a harsh and severe winter will occur.

Sea ice exported from the central Arctic Ocean is sensitive to large-scale atmospheric circulation such as AO due to its significant sink for ice. Over the past few decades, years of ice covering 3−4 m thick in the central Arctic Ocean has been replaced by an increas‐ing number of years of ice with a thickness of only 1 m, and an increasing number of ice-free zones have emerged (Wadhams, 2016). There may be more and more snow ice and frozen ice formation on the sea ice output channels in the future, which are not consid‐ered in the current sea ice models(Boyet al.,2019).

Studies on summer weather modes have found that polar cryosphere melting has a different contribu‐tion to heat waves, droughts, wildfires, and floods in the Northern Hemisphere during the summertime(Francis and Vavrus, 2012; Guoet al., 2014; Tanget al., 2014; Petrieet al., 2015; Mannet al., 2017;Vavruset al., 2017; Francis, 2018;Yorket al., 2018).Recent research suggests that the high-temperature heat waves in the summer since 1970s have frequent‐ly emerged with an increasing trend in Europe, main‐ly because the decrease in the the Arctic sea ice inten‐sity and snow cover area in Eurasia has weakened the temperature gradient between the Arctic and low- and mid- latitudes and affected mid-latitude jet stream and transient eddy activities (Zhanget al., 2020). These dynamic and thermodynamic circulations increase the likelihood of more persistent European blocking events that favor frequent and strengthened heat waves. Integrated analyses of the effects of Greenland sea ice, Eurasian snow cover, and the El Niño−South‐ern Oscillation (ENSO) on the Indian summer mon‐soon (ISM) and Korean summer monsoon (KSM) re‐vealed that Indian and Korean summer rainfalls showed nonlinear responses to ENSO and Greenland sea ice forcing, and variability in Eurasian snow pat‐terns may play a crucial role in ISM and KSM (Kimet al., 2020). The relationship between the late spring(May) snow water equivalent (SWE) over Siberia and the summer (July −August) rainfall in South-Central China show that anomalously low SWE over Siberia is robustly linked to summer rainfall in South-Central China(Shenet al.,2020).

Wildfire is one of the primary sources of BC de‐posited in the Arctic cryosphere besides gas flaring,shipping, and household burning (Hall and Loboda,2017). Studies on the relationship between wildfire and climate warming show that the prolonged snowfree season has led to shrub growth at the edge of for‐ests (Derksenet al., 2015), thereby increasing the oc‐currence frequency of wildfires in the boreal forest(Mårdet al., 2017). With an increase in summer tem‐perature, the fire probability rises dramatically. More wildfires mean more BC from the boreal forest,hence higher contribution to the Arctic warming. The rela‐tionship between wildfire and climate have shown possible thresholds: above an average July tempera‐ture of 13.4°C and below an annual moisture availabil‐ity of approximately 150 mm (Jollyet al., 2015;Younget al.,2016).The Arctic teleconnection and the interactions among regional feedback processes can lead to dry heat extreme events at mid-latitudes(Coumouet al.,2018).

5 Summary

Nowadays there is still a great deal of uncertainty about how cryosphere change affects weather and cli‐mate through different atmospheric circulation pro‐cesses at different spatial and temporal scales due to insufficient research, inadequate observational data and short time sequence, and imperfect models. For example, existing observation results support the con‐tribution of AA to colder wintersat mid-latitudes. Al‐though some model simulations also support observa‐tional evidence,most of the results do not show an as‐sociation between the Arctic change and the mid-lati‐tude weather, or there is excess heat exporting from the Arctic to lower latitudes (Cohenet al., 2020). A simulated test with or without AA indicates that the climatic impacts of AA are probably small outside the high latitudes. Thus, caution is needed to attribute mid-latitude changes to AA and sea ice loss on the ba‐sis of statistical analyses that cannot distinguish the im‐pact of AA from other correlated changes (Dai and Song, 2020). The differences between the models and observational studies, and even the contradictions among the models, make the understanding of the cli‐mate effect of the high-latitude cryosphere melting still ambiguous. Further research, continuous and improved observation, and synergetic modeling are needed to wholly improve the level of scientific understanding.

Understanding rapid warming and the climate feedback effect of accelerated melting of cryosphere is hindered by the current computing power and knowledge deficiencies on the impact of atmospheric dynamic processes(Francis,2018).How to"dance to‐gether" in harmony with the melting cryosphere re‐quires accurate identification of the processes be‐tween it and other changes and natural fluctuations,and the determination of the combined effects of vari‐ous changes and processes to solve a series of system‐atic related scientific problems. Researches in the fu‐ture have a long way to go. However, the changes in remote and cold Arctic region, increasingly affecting us, are closely related to the middle and low latitudes of dense population. Thus, in recent years, a large number of researches have involved in this region,showing a"hundred flowers bloom"landscape.In sum‐mary,the rapidly melting cryosphere's impact on glob‐al weather and climate is positive, and the magnitude of its feedback role needs to be further understood,es‐pecially its formation mechanism and reliability(Vavrus, 2018). The focus and controversy on this is‐sue will stimulate the study of the interaction between the cryosphere and atmosphere, further broaden the understanding the increasingly accelerated melting cryosphere's climate impact, and promote scientific consensus.

Acknowledgments:

This work is supported by the National Natural Sci‐ence Foundation of China (41730751, 41421061) and the Strategic Priority Research Program of the Chi‐nese Academy of Sciences, No. XDA23060700. The authors thank two anonymous reviewers for their sug‐gestions leading to significant improvement in the pa‐per.We thank all persons for their help to us.