Global Snow- and Ice-Related Disaster Risk: A Review

Over the last few decades, global warming has led to widespread shrinking of the cryosphere, with mass loss from ice sheets and glaciers (very high confidence); reductions in snow cover (high confidence) and Arctic sea ice, river, lake extent, and thickness (very high confidence); and increased permafrost temperature (very high confidence). Ice sheets and glaciers worldwide have lost mass (very high confidence) (IPCC 2021). The rapid cryosphere change has also a wide range and profound negative impact on economic and social system as well as its sustainable development, endangering human well-being and health.

From 1961 to 2016, global glacier mass cumulated to $-\mathrm{9,625}\pm \mathrm{7,975}\mathrm{Gt}$ ($1\mathrm{Gt}={10}^9\mathrm{t}$), corresponding to a contribution of $27\pm 22\mathrm{mm}$ to sea level, or a contribution of $0.5\pm 0.4\mathrm{mm}/\mathrm{year}$ when a linear rate is assumed. Since 1961, global annual mass loss reached 181 Gt, in which global annual mass loss continuously increased to 366 Gt during the latest pentad (2011–2016) (Zemp et al. 2019) [Fig. 1(a)]. From the mid-1920s to the beginning of this century, the extent of the northern hemisphere snow cover has been decreasing, especially in the 1980s. During May of 1965–2020, the extent reduction rate of snow cover reached $0.8\mathrm{million}{\mathrm{km}}^2/\mathrm{decade}$. In May 2020, snow cover extent in the northern hemisphere was 2.36 million ${\mathrm{km}}^2$, which was below the 1981–2010 average ($19.02\mathrm{million}{\mathrm{km}}^2$), with a sixth-smallest snow cover extent of May in the 54-year record. In the March–May season, the snow cover extent in the northern hemisphere, as recorded 1.95 million ${\mathrm{km}}^2$ below average, was the fourth-smallest value [Fig. 1(b)]. Under an RCP8.5 scenario, snowmelt is estimated to accelerate throughout the 21st century (AMAP 2017). Snow cover extent declined at an average rate of about $-12\%/\mathrm{decade}$ across a north–south transect of approximately 2,500 km (18–40°S) over the period 1986–2018 (Cordero et al. 2019).

Since 1979, the maximum extent of Arctic sea ice has continued to decline, with a decrease rate of $0.42\mathrm{million}{\mathrm{km}}^2/\mathrm{decade}$. Artic winter sea ice maximums in 2015, 2016, 2017, and 2018 were recorded at low levels, and the 12 lowest minimum extents from satellite image occurred in the last 12 years (NSIDC 2022) [Fig. 1(c)]. The volume of Arctic sea ice in September has declined by 75% since 1979 (Overland et al. 2019). In contrast to the Arctic, the area of sea ice showed a small positive increase ($1.6\%\pm 0.4\%$ per decade between 1979 and 2016) in the Southern Ocean. However, extreme changes have occurred in recent years—the extent of Antarctic sea ice reached maxima in three successive winters (2012, 2013, and 2014), followed by a minimum record in the summertime of March 2017 (Shepherd et al. 2018). Since the 1970s, the permafrost temperature has risen between 0.50°C and 2.00°C in most parts of the Arctic (Romanovsky et al. 2010). In the reference decade 2007–2016, the ground temperature increased by $0.39°\mathrm{C}\pm 0.15°\mathrm{C}$ around the depth of zero annual amplitude of the successive permafrost area. The higher the latitude locates, the lower the ground temperature and the higher the warming amplitude. Taking 1,113, 117, and 356 boreholes in the Arctic regions of Russia, Alaska, and Canada as examples, the permafrost temperature increased by 0.42°C (1992–2017), 0.66°C (1978–2016), and 0.38°C per decade (1978–2014), respectively (GTN-P 2015) [Figs. 1(d–f)]. Similar to the continued warming of the Arctic permafrost, the Antarctic permafrost warmed by $0.37°\mathrm{C}\pm 0.10°\mathrm{C}/\mathrm{decade}$ between 2008 and 2016 (Biskaborn et al. 2019).

As a destructive event or phenomena, SIRD changes in the cryosphere have a great effect on humans or their environment for survival. Its formation requires not only cryosphere hazards as incentive conditions, but also the objects that are subject to disasters such as people, property, resources, and infrastructure. Generally, a SIRD is the result of combined effects of the cryosphere hazard, the disaster’s formative environment, and the vulnerability of the economic and social system (the ability to prevent and reduce disasters). SIRDs include ice/snow avalanches, glacial surges, glacial–snow flood/debris flows, melting snow floods, freeze–thawing, ice jam/floods, iceberg sea ice, coastal freeze–thaw erosion, snowstorms, blowing snow, freezing rain and snow, hail, frost freezing, and other disasters according to their trigger factors (Fig. 2).

Cryosphere hazards are the result of the imbalance of the cryosphere and its threshold can be identified as a key indicator of the occurrence of cryosphere events (or snow- and ice-related events). The disaster’s formative environments include natural factors such as climate, topography, geological activity (e.g., earthquake and volcanic eruption), ecological vegetation, elevation, river distribution, and social factors such as population, economy, and infrastructure. For cryosphere events in the same intensity, the higher the sensitivity of the disaster’s formative environment, the greater the probability of disaster and the greater the losses. In addition, the size of the damage depends on the exposure and vulnerability of the disaster-bearing bodies. Exposure refers to the number of assets affected by cryosphere hazards, and the vulnerability is the tendency of the exposure elements to be affected by cryosphere hazards If the disaster-bearing body is a community or social system, the vulnerability also includes the size of the disaster prevention and mitigation capabilities of the system. The greater the vulnerability, the greater the risk of the SIRD. Sometimes, the SIRD size is determined by the scale of the hazard, and sometimes by the degree of exposure or vulnerability as well as the disaster’s formative environment.

The occurrence and impact of different types of SIRDs also need time and spatial scale. Some disasters occur instantaneously, while others need a longer time. Some disasters may be very local, while others have valley, watershed, regional, and even larger temporal scales. Among them, snow/ice avalanches are locally instantaneous in mountainous areas. Glacial debris flows, GLOFs, and glacial surges occur on an hourly scale at valley scale. Frost and iceberg disasters occur on a daily scale and spatially on a regional scale. Snowstorms, low-temperature rain and snow, and freeze–thaw disasters occur on a daily or monthly scale and mostly on a large regional scale. There are seasonal, interannual, or interdecadal scales of freezing and thawing erosion along the Arctic coastal zone. SLR disasters caused by cryosphere and formative disasters have an interdecadal and even longer time with global spatial scales (Wang and Xiao 2019) (Fig. S1).

SIRDs that widely distributed in middle–low latitude, high-altitude mountainous areas, and high latitude areas have significant economic and social impacts on humans, where the level of economic development is relatively low, and the ability of disaster prevention is limited due to the inconvenient accessibility. Global SIRD sites have good spatial correlation with the significant warming areas and the areas that frequently experience extreme weather and climate events. Among them, areas exposed to rapid cryosphere change with an undoubtedly high risk and huge potential impact need to be carefully concerned. Snow/ice avalanches often occur in the middle- and high-latitude snow-covered mountainous regions throughout the world. Glacial surges mainly occur in the Karakorum and Pamirs, Tianshan, the Nordic Svalbard Islands, Nova Scotia, Iceland, Greenland, Alaska and Yukon in North America, the Canadian Arctic, and the Andes (Mukherjee et al. 2017; RGI 2017) and often cause GLOFs (Bazaiaf et al. 2021). Mongolia, the Caucasus, Georgia, and Western China are frequently and severely affected by snow disasters in pastoral areas (Wang et al. 2019a), where they are often accompanied by blowing snow. Glacial, ice, and snow floods are widely distributed in the high-altitude mountains of mid–high latitude regions in Russia, the Caucasus, Central Asia, the northwestern coastal mountains of the United States and Canada, the Himalayas, and the Karakorum Mountains. GLOFs mainly occur in the Hindu Kush-Himalayan Mountains, the southwest coast of Canada, the Andes Mountains in South America, and the Qing-Tibetan Plateau (QTP). Lahars or floods caused by ice-clad volcano eruption mostly appear at Mexico, Colombia, Ecuador, Chile, and Iceland (Granadosa et al. 2021). Freeze–thaw disasters mainly occur in Arctic countries, the QTP, and northeastern China, and they often endanger the infrastructure and construction of road networks, oil/gas pipelines, airports, and residential areas (Hjort et al. 2018). Iceberg disasters are mainly concentrated in the ocean areas off Antarctic coast, surrounding Greenland and northeastern Canada. Sea ice disasters focus on coastal and offshore areas of Arctic countries, especially in the Arctic shipping routes (Kubat et al. 2012, 2015). Coastal freeze–thaw erosion mainly occurs on the northern coast of Russia, Alaska, and Canada. Globally, the SLR under the influence of the cryosphere mainly threatens the islands and coasts in the mid–low latitudes. SLR often overlaps with storm surges to form coastal floods, causing coastal erosion and directly affecting social systems. Snowstorm disasters mainly appear in the northeastern United States, southwestern Canada, Siberia, northeastern China, Altai, Europe, and Japan, often endangering aviation and transportation networks. Freezing rain and snow disasters mainly occur in southern China. Hail disasters are mainly distributed along the west coast of the midlatitudes (from British Columbia in Canada to the Californian coast in the United States, Western Europe, the Nordic coast, the Mediterranean coast, and northwestern Japan) and in some alpine areas (Fig. 3).

On land, SIRDs mostly occur in middle–low latitude areas and in high-latitude permafrost regions and result in more frequent and more serious disasters such as snowstorm disasters, freezing rain, hail, frost disasters, avalanches, glacier surges, GLOFs, glacier flood/debris flow, melting snow flood, ice jam/flood, and freeze–thaw disasters (Fig. S3 and Table S1).

Snowstorms often result in traffic slowdowns or interruptions, highway closures, and airport flight delays or cancellations. From 1949 to 2000, snowstorms in the United States caused a total of $21.6 billion in damages, with losses increasing since 1980, largely due to population growth and property fragility (Changnon and Changnon 2006). During the period from January to February 2012, damaging winter snowstorms occurred in Japan, accompanied by an avalanche. The incident resulted in 83 deaths, several bridge collapses, and the blocking of highways and roads. During the same period, a winter cold wave caused damage in eastern, southern, and western Europe, snowdrifts up to 8-m deep, extreme frost, and low temperatures ($-39°\mathrm{C}$) resulted in 541 deaths and direct economic losses of $850 million. In November 2018, more than 1,600 flights at US airports were cancelled and another 15,000 were delayed due to a snowstorm. Snow drifting (or blowing snow and blizzard) is an aeolian process that has been widely documented at high latitudes, high altitudes, and in areas with complex topography (Nishimura et al. 2014). Snow can block traffic lines, cause power outages, and have serious impacts on economic and social systems. Between 1959 and 2014, there were 713 drifting snow events in contiguous US regions. However, blizzard-related casualties have decreased due to improved forecasting, warning, and communication mechanisms (Coleman and Schwartz 2017).

The occurrence of freezing rain and snow events is often characterized by low temperature, high humidity, and a low wind speed. The occurrence of such disasters is caused by the accumulation of multiple continuous rainfall/snow weather processes. The disasters are more widespread and the loss is more serious. During the 1982–1994 period, the United States experienced an average of 16 freezing rain and snow events per year, which was much higher than the number of snowstorms. In eastern Canada, it has been projected that the freezing rain frequency will increase during December to February. At the beginning of 2012, Europe experienced persistent extreme low temperatures. Rare freezing rain–snow events and snowfall were observed in many areas along the northern coast of the Mediterranean Sea. More than 800 people died as a result of the disaster in Eastern Europe alone (Blunden et al. 2013). In 2008, China’s freezing rain and snow events caused serious losses, with 107 deaths and direct economic losses amounted to 111.9 billion CNY (Cheng et al. 2011).

Hailstones associated with thunderstorms is a destructive phenomenon that occurs primarily in some parts of the eastern and central United States, Europe, and Central Africa (Allen et al. 2019). In the United States, average annual property damage has begun to exceed $10 billion, with severe hail events affecting large cities typically hitting $1 billion in damages (Gunturi and Tippett 2017), although these losses do not include impacts on agriculture. In May 2017, Colorado experienced the most expensive hailstorm in its history, resulting in 267,000 claims and $2.3 billion in damages in the Denver area, according to the Rocky Mountain Insurance Information Association. Then, in June 2018, waves of vicious hailstorms swept the Front Range: the first hailstorm hit Fountain, causing $169 million in damage (Brendza 2019). From April 17 to May 10, 2017, much of Europe was hit by a cold snap with overnight frosts. Due to an unusually warm spring, the germination process has advanced, and losses are at historic levels—especially for fruit and wine growers: economic losses are estimated at EUR 3.37 billion, of which about EUR 600 million are insured (Faust and Herbold 2018). In particular, super cooled droplets and high-density ice crystals are dangerous for civil aviation (Kjellsson 2015; Bernabò et al. 2015).

Frost occurs when the surface or vegetation temperature drops to 273 K, and the plants are frostbitten. Frost disasters have great spatial heterogeneity. Frost remains a major threat to agricultural productivity in the highlands of East Africa due to frost damage (Kotikot et al. 2020). Frost accounted for 34.5% of total insured crop losses in Greece in 1999–2011 (Papagiannaki et al. 2014). In Central Europe, studies have shown that frost occurs earlier than the phenology period, so there may have been a downward trend in frost damage to plants from the 1950s to the late 1990s (Rigby and Porporato 2008). The number of frost days is clearly decreasing in Asia (Dong et al. 2018). In China, the annual averaged frost weather, more than 180 days, mainly occurs in the northeastern part of the QTP, the Tianshan Mountains, and the Xingan Mountains. The number of frost days exhibits a linear increasing trend in northern Central China and the QTP, and a decreasing trend in the Yangtze River basin and to the south. There is an evident increasing trend in the nationwide average number of frost days, with a rate of increase of $2.03\mathrm{days}/\mathrm{decade}$, so the frequency of frost days is increasing (Zhang et al. 2015). In April 2018, a large-scale flowering frost disaster reduced pear production by 60%–80% throughout northern and northwestern China and some pear gardens were destroyed (Zhang et al. 2018).

The occurrence of snow/ice avalanches is sudden with a short duration. Between 2000 and 2010, the avalanches in Europe and North America killed approximately 1,900 people. In 2010, 2012, and 2016, three large-scale avalanches occurred in the Hindu Kush Afghanistan and Kashmir region killing 419 people (Wang and Ren 2012). Two million people in Afghanistan were exposed to avalanches and more than 153,000 people were affected between 2000 and 2015 (Federica et al. 2017). In 2017, there were more than 250 avalanches in the Swiss Alps, which killed 26 people. In 2002, a large-scale ice avalanche occurred in the Kolka Glacier in northern Ossetia, Russia, and 130 villagers were killed. Between 2015 and 2018, glacial surge disasters at Karayaylak glacier in the East Pamir Plateau and ice avalanche disasters at Aruco glacier in 2016, Qinglonggou glacier in 2016, and the Sedongpugou glacier in 2018 on the QTP have caused a certain amount of losses (Lü et al. 2016; Tian et al. 2017; Kääb et al. 2018; Chen et al. 2020). In particular, on February 7, 2021, a catastrophically large glacial and rock avalanche occurred in the Ronti Gad, Rishiganga, and Dhauliganga valleys in the Chamoli region of Uttarakhand, India, causing widespread damage and severely damaging two hydropower projects. Over 200 people were killed or are missing (Shugar et al. 2021; Thayyen et al. 2021). Snow disasters greatly affect the sustainable development of pastoral areas around the world. In 2004, 8.5 million livestock in Mongolia died because of snow disasters. During 2009–2017, more than 10.93 million animals in Mongolia died in a snow disaster (Fernandez-Gimenez et al. 2012). In the last 60 years (1951–2015), 238 snow disasters were recorded above the scale of the QTP, with 12 million dead animals (Wang et al. 2019a).

Melting glaciers, snow, and ice also contribute to increased risks of floods and mudslides (Motschmann et al. 2020). In the last 30 years, the frequency and impact of GLOFs and snow–ice floods in central Asia have increased significantly. Nearly $200{\mathrm{km}}^2$ of farmland was threatened by the aforementioned floods in the Aksu River irrigation district of Tianshan, China. In March 2010, snowmelt floods occurred in Xinjiang, western China, affecting more than 1.3 million people, killing 13 people, and causing economic losses of more than $20 billion (Ding et al. 2021). In 2015, higher temperatures increased the rate of snow and ice melting and caused widespread flooding in east-central Kazakhstan, with 35 people rescued and 14,790 evacuated (ECMIAK 2015). In May 2018, thousands of people in British Columbia, Canada, were forced to evacuate their homes due to snowmelt flooding (Rokaya et al. 2018). In particular, the frequency of floods in the lower Indus, Ganges, and Brahmaputra are projected to rise significantly in the 21st century, putting the livelihoods of 220 million people at risk (Veh et al. 2020). The interaction between volcanic activity and ice and snow bodies can also lead to severe disasters such as floods/landslides. Volcanic activity interacted with ice bodies during recent volcanic crises: Popocate´petl, Mexico; Nevada del Huera, Colombia; and Yaima and Villari card, Chile (Cranados et al. 2015).

There have been 1,348 recorded glacial lake outburst events worldwide. GLOF disasters have directly caused at least 13,160 deaths in the world (Carrivick and Tweed 2016; Wang and Zhou 2017; Kougkoulos et al. 2018). Especially, in 2013, a GLOF event in the Chorabari glacier valley left behind a death toll of more than 5,000 people (Shukla and Sen 2021). Currently, glacial lakes hold about 0.43 mm of sea level equivalent (Shugar et al. 2021). Ice jam is a blockage of river flow that causes a temporary rise in water levels, which often results in higher water levels and more extensive damage than open water events (Lindenschmidt et al. 2018), and poses a serious threat to riverine communities. In the northern hemisphere, nearly 60% of rivers experience significant seasonal effects of river ice (Frolova et al. 2015). In Europe, ice jams were observed along the southern Baltic coast during the winters of 1995, 2003, 2010, and 2011 and affected narrower and lower beaches following ice melt in late spring (Ryabchuk et al. 2011). In Russia, the 2001 ice-jam flood on the Lensk River caused a loss of $103 million (Kilmyaninov 2001), and the 2010 ice-jam flood on the Lena River damaged 2,000 houses, and caused a loss of 1.2 billion rubles including the loss of more than 2,300 livestock (Takakura 2016). In Canada, ice-jam flooding is also common in Quebec and the Northwest Territories (Burrell et al. 2015; Rokaya et al. 2018). In Alaska, annual ice-jam flood losses are estimated at $138 million (White et al. 2007; Carr et al. 2015). In China, the Yellow River often suffers from ice-jam flood disasters, causing huge casualties and property losses (Ye et al. 1999; Fu et al. 2014).

Degradation of near-surface permafrost can pose a serious threat to the utilization of natural resources and sustainable development to the Arctic and QTP’s communities (Hjort et al. 2018) and result in significant impacts on the stability of engineering structures in cold areas (Wu and Niu 2013; Wang et al. 2020). From 1984 to 2015, the frequency of Arctic thaw slump showed a clear increasing trend, and climate warming is expected to continue to cause a large number of thaw slumps (Lewkowicz and Way 2019). Changes in the permafrost will exacerbate the coping costs to adapt these negative impacts. Since the 1970s, the degradation of permafrost has led to a decline in the ground-supporting infrastructure in five regions of Russia: Chukotka AO, Republic of Sakha, Taymyr AO, Yamalo-Nenets AO, and Nenets AO (Streletskiy et al. 2012). In areas rich in ice, the permafrost degraded or even disappeared, causing significant subgrade subsidence and instability of the bridge foundation. The rapid degradation of permafrost means that the cost of disaster prevention and mitigation for various types of engineering buildings in permafrost regions will increase further.

Especially, further warming is projected to further expand permafrost thaw and loss of seasonal snow cover, land ice, and Arctic sea ice (high confidence) (IPCC 2019). Reductions in ocean ice and snow cover lead to a decrease in albedo, and thawing permafrost accelerates methane emissions, warming the planet and exacerbating global warming (IPCC 2018). Emissions of greenhouse gases and albedo changes from Arctic melting are predicted to more than double the Arctic’s contribution to global warming by 2100. These effects are slow but global. This may need to be viewed, considered, and handled as a global catastrophe.

In the ocean, icebergs, floating ice movement, freezing and ablation of sea ice, coastal freeze–thaw erosion, and SLR disasters often affect Arctic coastal assets and offshore facilities, maritime activities, and shipping safety, such as navigation channel congestion, increasing the risk to vessel and navigation safety, and damage to ports, offshore drilling platforms, coastal engineering, offshore aquaculture, and activities (hunting, travel) of offshore residents (Table S2).

Icebergs are large pieces of glacial ice that float in bodies of water and their collisions mainly occur in Alaska, as well as in Baffin, the Gulf, the Hudson Strait, and the Labrador Sea. Decreased visibility due to fog, rain, snow, and night is the main cause of iceberg disasters. Iceberg disaster records extend from the year 1619, in Svalbard, to the sinking of the Titanic in 1912, to the MS Hans Hedtoft disaster in 1959. Most incidents occurred in the 19th and 20th centuries, with collisions with icebergs steadily increasing with maritime trade, with 26% of incidents resulting in shipwrecks or abandonment (Bigg et al. 2018). At times, the grounding of icebergs can lead to the formation of scour and pose a hazard to pipelines and subsea installations (King et al. 2003). From 1980 to 2005, 57 accidents occurred in the northern hemisphere involving icebergs, with an accident rate of 2.3 per year. In particular, on November 23, 2007, the US Explorer (MS Explorer) hit an iceberg in the Antarctic water zone and sank. Although no one was killed, the accident was still named “The Modern Titanic Incident” by The New York Times (Hill 2010).

The presence of sea ice (especially floating ice) in northern regions has significant economic, environmental, and social implications (Kubat et al. 2012, 2015). Sea ice is the main load of offshore and shore-based facilities and ships in ice-prone areas, threatening the safe operation of these facilities and ships. Under the dynamic interaction between sea ice and ocean structure, sea ice will cause different failure modes, including plastic failure, cracking failure, and brittle failure (Newman 1971). Sea ice undergoes different fragmentation modes at different strain rates. Sea ice mainly presents linear elastic brittle failure under high strain rate load, while ice mainly presents elastoplastic failure under low strain rate load (Hu et al. 2016). Seismic dynamics further complicate this influence process (Huang et al. 2021a, b, c). In extreme cases, sea routes can be blocked and fractured, ships can be damaged, and smaller ships can be lifted completely onto the ice. Ships recorded incidents plagued by floating ice in the Beaufort Sea, Peel Sound, Baffin Bay, Han Bay, the Labrador Coast, the east coast of Newfoundland in Canada, the Great Lakes in Alaska in the United States, western Greenland, Europe, and part of the Arctic (Lilover and Kõuts 2012; Mironov et al. 2012; Lensu et al. 2013). For example, in 2007, strong southwesterly winds, ocean currents, irregular shorelines, massive shoals, further drift of floating ice, and intrusion of multiyear ice caused damage to severe pressure buildup along the east coast of Newfoundland and Labrador in April and May. As a result, a large number of fishing vessels were trapped in the pressure ice field (Kubat et al. 2012, 2015). In 2010, in the southern Bohai Bay in China, a ship loaded with 1,000 t of fuel oil was squeezed by floating ice and the cabin was flooded.

Rapidly melting sea ice increases the accessibility to sea areas that were previously difficult to reach in the late spring and early fall, but it greatly affects coastal assets, offshore facilities, and maritime activities, the ecosystems that are highly dependent on sea ice and its habitats and on the production activities of the Arctic Indigenous Peoples. For example, Arctic bears are very sensitive to sea ice changes. As the range and thickness of sea ice continues to decrease, the sites or areas of habitats that Arctic bears depend on for their survival and hunting continue to shrink. The number of Arctic bears in the West Hudson River Bay is decreasing, and their reproductive success rate is declining (Regehr et al. 2007). If sea ice continues to melt, nearly two-thirds of polar bears will become extinct in the future (Durner et al. 2009). Sea ice changes also have a huge impact on coastal aquaculture due to great changes of salinity of seawater during its icing and melting process (Zhang et al. 2013). For instance, in the last 10 years, the fishery in the Bohai Rim region of China has become the industry most affected by sea ice disasters, accounting for more than 90% of losses (Wu et al. 2015). Since 2010, the direct economic losses caused by sea ice disasters in the Bohai Sea and the northern Yellow Sea in China have reached 7.7 billion CNY (Zuo et al. 2019). Especially, delayed sea ice exposes coastlines to erosion and enhances thermal subsidence of coastal permafrost, accelerating coastal retreat rates over time (Hajo and Andrew 2015).

The coastal freeze–thaw erosion and SLR are closely related to natural habitat, homeland security (i.e., losing land), and often threaten planned or existing Arctic coastal infrastructure, the viability of low-lying coastal communities and cultural heritage passed on from the early explorers and indigenous peoples (Fritz et al. 2017). Especially, it may lead to changes in coastal food webs by releasing large amounts of organic carbon and nutrients to nearshore zones (Günther et al. 2015). The coastal freeze–thaw erosion and its destruction is caused by the kinetic energy of waves and the thawing of permafrost in coastal areas (Barnhart et al. 2014). Arctic permafrost makes up 34% of Earth’s coasts undergoing large erosion, with its rate as high as $25\mathrm{m}/\mathrm{year}$ (Fritz et al. 2017). The annual loss of land due to coastal erosion in the Arctic region is approximately $51{\mathrm{km}}^2$ (Lantuit et al. 2012; Wright et al. 2019). For example, ice strongly eroded the coastal land and human structures in Neva Bay on the northern coast of the Baltic Sea in February 2008 (Ryabchuk et al. 2011). The US Government Accountability Office found that flooding and erosion affected 184 of 213 Alaska Native communities, of which 31 were threatened and 12 were planned for relocation (Bronen and Chapin 2013).

The greatest impact of the SLR, which is mainly caused by the cryosphere and ocean thermal expansion, is undoubtedly on the world’s small islands, coastal cities, deltas, and other low-lying coastal areas. A minimal SLR may significantly increase the frequency and intensity of coastal flooding. An SLR of 0.5 m will increase the frequency of coastal flooding 100 or even 1,000 times in many coastal areas. SLR has caused coastal systems and low-lying areas to be increasingly and adversely affected by coastal flooding and coastal erosion. Since the mid-19th century, SLR rates have been higher than the average rate over the last 2,000 years (high confidence). During the 1901–2010 period, average global SLR increased by 0.19 m (0.17–0.21 m) (IPCC 2013). Studies have shown that mean high water changes exceed $\pm 10\%$ of the SLR at 13 of 136 major coastal cities with 1 m of SLR (Pickering et al. 2017). Currently, more than 300 million people live in low-altitude coastal areas and suffer tens of billions of dollars in damage loss each year (Wahl et al. 2017). Based on a rough estimate, approximately 1.30% of the global population is exposed to a once-in-a-century flood extent (Muis et al. 2016). A current study estimates that globally a total of 110 million people live on land below the current high tide line and 250 million live on land below the annual flood level (Kulp and Strauss 2019). As a direct consequence of SLRs, the frequency of floods, which bring more than $ 9 trillion in goods and services to US coastal communities annually, has doubled since 2000, and now averages 2–6 times a year (Sweet et al. 2020). From 1980 to 2017, China’s coastal SLR showed a fluctuating upward trend, with an upward rate of $3.3\mathrm{mm}/\mathrm{year}$. SLR has a great impact on the densely populated areas of China’s coastal cities (Gao et al. 2019).

The risk assessment of the SIRD depends not only on the probability of the occurrence of the snow- and ice-related hazard, but also on the future estimation of the density and scale of the exposed elements. In the near-term (2035–2050), the loss of glacial mass, permafrost thawing, snow cover, and reductions in Arctic sea ice extent are expected to continue. The Greenland and Antarctic ice sheets are projected to experience increasing mass loss throughout the 21st century. In the future, due to population growth and improved socioeconomic development, cryospheric hazards are expected to continue to increase, especially in the Andes, the Asian High Mountains, the Caucasus, and the European Alps (IPCC 2019).

The reduction in future extreme cold events is statistically more significant than preindustrial times, and extreme cold events tend to be more sensitive to global warming than extreme warm events (Wang et al. 2019b). During the extreme cold weather, future hail damage is expected to increase due to increased population density and wealth (Prein and Holland 2018). Global warming may reduce the frequency of freezing rain events (Lambert and Hansen 2011), while the global average number of terrestrial frost days will also change significantly. The results showed that the frost days in each model showed a decreasing trend in the 21st century (Wang et al. 2017). Several studies suggest that the frequency of hailstones will increase and the frequency of small hailstones will decrease in North America in the coming decades (Trapp et al. 2019). Il Jeong and Sushama (2018) have predicted an increase in winter rain and snow events and a decrease in spring in North America during the 2041–2070 period (RCP4.5 and RCP8.5), as confirmed by Musselman et al. (2018). Evidence since the IPCC Fifth Assessment Report indicates that the frequency of snow and rain events is expected to increase at high altitudes, appearing earlier in spring and later in fall, and decreasing at lower altitudes (high confidence) (Hock et al. 2019). In the RCP4.5 and RCP8.5 scenarios, by the middle of the 21st century, the minimum daily temperature in China was 1.7°C and 2.2°C higher than it was during 1986–2005. The frost days decreased by 13 and 16 days, respectively, and the freezing days also decreased by 10 and 12 days, respectively (Qin 2018). Although the number of low temperature events shows a decreasing trend, it still cannot be ignored due to the significant impact of staged low temperature freezing rain and snow events.

Avalanches involving wet snow are expected to be more likely to occur in winter at all altitudes due to surface melt or rain and snow (Castebrunet et al. 2014). In areas and altitudes where snow cover is significantly reduced, the total number of avalanches is expected to be the same because swipe distance will decrease (Mock et al. 2017). Studies have also shown that warm temperatures in winter and early spring are conducive to the formation of moist snow, which in turn leads to an increase in the frequency of wet avalanches, which is likely to have a major impact on areas with frequent human activities, especially in steep sloped, subalpine areas (Ballesteros-Cánovas et al. 2018). There is also high confidence that the number and size of glacial lakes will continue to increase in most regions over the next few decades, and that new lakes will develop close to steep and potentially unstable lake–mountain walls that can more easily trigger the effects of eruptions snow, ice or rock collapse, landslides, earthquakes, and extreme precipitation (Allen et al. 2016; Haeberli et al. 2017; Sæmundsson et al. 2017; Wang et al. 2020). Over the last century, many glacial dammed lakes have experienced decades-long periodic or episodic outbursts. We expect that the global trend of glacial lake growth will continue, and may even accelerate, as glacial melting and retreat proceed, increasing the supply of vulnerable moraine lakes in a warming world (Shugar et al. 2021). Extensive degradation of near-surface permafrost is predicted to have detrimental effects on northern communities, ecosystems, and engineered systems in the 21st century (Teufel and Sushama 2019). Compared with 1986–2005, the permafrost area near the surface of the northern hemisphere will be reduced by $37\%\pm 11\%$ (RCP2.6), $51\%\pm 13\%$ (RCP4.5), $58\%\pm 13\%$ (RCP6.0), and $81\%\pm 12\%$ (RCP8.5) (Collins et al. 2013) by 2080–2099. Among them, Russia’s permafrost area will decrease by $32\%\pm 11\%$ (RCP2.6), $49\%\pm 13\%$ (RCP4.5), $55\%\pm 14\%$ (RCP6.0), and $76\%\pm 12\%$ by the end of the century (RCP8.5) (Streletskiy et al. 2015). This is highly likely to affect the maintenance of the planned Beijing–Moscow high-speed railway and the oil and gas pipelines on the Yamal Peninsula.

Future iceberg and meltwater discharges from the Antarctic ice sheet could significantly exceed current levels, with strong implications for future climate and sea levels (Schloesser et al. 2019). As polar warming increases, polar regions are opening up to more human activity as sea ice shrinks, but the number of icebergs is more likely to increase (Barnes et al. 2014), which will result in likely related iceberg disasters increasing in the future. According to estimates, the polar annual average coastal erosion rate is about $0.5\mathrm{m}/\mathrm{year}$, and thus, the annual loss of land area reaches $51{\mathrm{km}}^2$ in the Arctic region due to coastal erosion (AWI 2013). Central estimates in the recent literature broadly agree that global mean sea level is likely to rise 20–30 cm by 2050 (Dangendorf et al. 2017; Kulp and Strauss 2019). In the case of SLR, the risk of damage may increase significantly. Without effective adaptation measures, by the end of the century, the potential damage may reach 10% of the global GDP (Wahl et al. 2017). Global climate models consistently predict further reductions in Arctic sea ice (IPCC 2019). Even under the RCP8.5 scenario, much of the Arctic Ocean would still freeze in winter, but this seasonal sea ice is expected to be thinner, more mobile, and more dispersed (AMAP 2017), leaving a significant potential risk for shipping safety.

Numerical modeling is an alternative method to detect cryosphere events trends using observation records. However, some disasters in the cryosphere are low-frequency disasters, and the sample size of the disasters is insufficient. However, with the further warming of the climate, the cryosphere events will continue to increase. Against this background, the probability of occurrence of SIRDs will increase significantly.