Editing IPCC:AR6/SROCC/Chapter-2 (section)

==== 2.3.2.1 Observed and Projected Changes ====

<div id="section-2-3-2-1observed-and-projected-changes-block-1"></div>

<span id="unstable-slopes-landslides-and-glacier-instabilities"></span>
===== 2.3.2.1.1 Unstable slopes, landslides and glacier instabilities =====

Permafrost degradation and thaw as well as increased water flow into frozen slopes can increase the rate of movement of frozen debris bodies and lower their surface due to loss of ground ice (subsidence). Such processes affected engineered structures such as buildings, hazard protection structures, roads, or rail lines in all high mountains during recent decades (Section 2.3.4). Movement of frozen slopes and ground subsidence/heave are strongly related to ground temperature, ice content, and water input (Wirz et al., 2016 <sup>[[#fn:r447|447]]</sup> ; Kenner et al., 2017 <sup>[[#fn:r448|448]]</sup> ). Where massive ground ice gets exposed, retrogressive thaw erosion develops (Niu et al., 2012 <sup>[[#fn:r449|449]]</sup> ). The creep of rock glaciers (frozen debris tongues that slowly deform under gravity) is in principle expected to accelerate in response to rising ground temperatures, until substantial volumetric ice contents have melted out (Kääb et al., 2007 <sup>[[#fn:r450|450]]</sup> ; Arenson et al., 2015a <sup>[[#fn:r451|451]]</sup> ). As documented for instance for sites in the European Alps and Scandinavia for recent years to decades, rock glaciers replenished debris flow starting zones at their fronts, so that the intensified material supply associated with accelerated movement (Section 2.2.4) contributed to increased debris flow activity (higher frequency, larger magnitudes) or slope destabilisation (Stoffel and Graf, 2015 <sup>[[#fn:r452|452]]</sup> ; Wirz et al., 2016 <sup>[[#fn:r453|453]]</sup> ; Kummert et al., 2017 <sup>[[#fn:r454|454]]</sup> ; Eriksen et al., 2018 <sup>[[#fn:r455|455]]</sup> ).

There is ''high confidence'' that the frequency of rocks detaching and falling from steep slopes (rock fall) has increased within zones of degrading permafrost over the past half-century, for instance in high mountains in North America, New Zealand, and Europe (Allen et al., 2011 <sup>[[#fn:r456|456]]</sup> ; Ravanel and Deline, 2011 <sup>[[#fn:r457|457]]</sup> ; Fischer et al., 2012 <sup>[[#fn:r458|458]]</sup> ; Coe et al., 2017 <sup>[[#fn:r459|459]]</sup> ). Compared to the SREX and AR5 reports, the confidence in this finding increased. Available field evidence agrees with theoretical considerations and calculations that permafrost thaw increases the likelihood of rock fall (and also rock avalanches, which have larger volumes compared to rock falls) (Gruber and Haeberli, 2007 <sup>[[#fn:r460|460]]</sup> ; Krautblatter et al., 2013 <sup>[[#fn:r461|461]]</sup> ). These conclusions are also supported by observed ice in the detachment zone of previous events in North America, Iceland and Europe (Geertsema et al., 2006 <sup>[[#fn:r462|462]]</sup> ; Phillips et al., 2017 <sup>[[#fn:r463|463]]</sup> ; Sæmundsson et al., 2018 <sup>[[#fn:r464|464]]</sup> ). Summer heat waves have in recent years triggered rock instability with delays of only a few days or weeks in the European Alps (Allen and Huggel, 2013 <sup>[[#fn:r465|465]]</sup> ; Ravanel et al., 2017 <sup>[[#fn:r466|466]]</sup> ). This is in line with theoretical considerations about fast thaw of ice filled frozen fractures in bedrock (Hasler et al., 2011 <sup>[[#fn:r467|467]]</sup> ) and other climate impacts on rock stability, such as from large temperature variations (Luethi et al., 2015 <sup>[[#fn:r468|468]]</sup> ). Similarly, permafrost thaw increased the frequency and volumes of landslides from frozen sediments in many mountain regions in recent decades (Wei et al., 2006 <sup>[[#fn:r469|469]]</sup> ; Ravanel et al., 2010 <sup>[[#fn:r470|470]]</sup> ; Lacelle et al., 2015 <sup>[[#fn:r471|471]]</sup> ). At lower elevations in the French Alps, though, climate driven changes such as a reduction in number of freezing days are projected to lead to a reduction in debris flows (Jomelli et al., 2009 <sup>[[#fn:r472|472]]</sup> ).

A range of slope instability types was found to be connected to glacier retreat (Allen et al., 2011 <sup>[[#fn:r473|473]]</sup> ; Evans and Delaney, 2015 <sup>[[#fn:r474|474]]</sup> ). Debris left behind by retreating glaciers (moraines) slid or collapsed, or formed fast flowing water-debris mixtures (debris flows) in recent decades, for instance in the European and New Zealand Alps (Zimmermann and Haeberli, 1992 <sup>[[#fn:r475|475]]</sup> ; Blair, 1994 <sup>[[#fn:r476|476]]</sup> ; Curry et al., 2006 <sup>[[#fn:r477|477]]</sup> ; Eichel et al., 2018 <sup>[[#fn:r478|478]]</sup> ). Over decades to millennia, or even longer, rock slopes adjacent to or formerly covered glaciers, became unstable and in some cases, eventually collapsed. Related landslide activity increased in recently deglacierised zones in most high mountains (Korup et al., 2012 <sup>[[#fn:r479|479]]</sup> ; McColl, 2012 <sup>[[#fn:r480|480]]</sup> ; Deline et al., 2015 <sup>[[#fn:r481|481]]</sup> ; Kos et al., 2016 <sup>[[#fn:r482|482]]</sup> ; Serrano et al., 2018 <sup>[[#fn:r483|483]]</sup> ). For example, according to Cloutier et al. (2017) <sup>[[#fn:r484|484]]</sup> more than two-thirds of the large landslides that occurred in Northern British Columbia between 1973–2003, occurred on cirque walls that have been exposed after glacier retreat from the mid-19th century on. Ice-rich permafrost environments following glacial retreat enhanced slope mass movements (Oliva and Ruiz-Fernández, 2015 <sup>[[#fn:r485|485]]</sup> ). At lower elevations, re-vegetation and rise of tree limit are able to stabilise shallow slope instabilities (Curry et al., 2006 <sup>[[#fn:r486|486]]</sup> ). Overall, there is ''high confidence'' that glacier retreat in general has in most high mountains destabilised adjacent debris and rock slopes over time scales from years to millennia, but robust statistics about current trends in this development are lacking. This finding reconfirms, and for some processes increases confidence in related findings from the SREX and AR5 reports.

Ice break-off and subsequent ice avalanches are natural processes at steep glacier fronts. How climate driven changes in geometry and thermal regime of such glaciers influenced ice avalanche hazards over years to decades depended strongly on local conditions, as shown for the European Alps (Fischer et al., 2013 <sup>[[#fn:r487|487]]</sup> ; Faillettaz et al., 2015 <sup>[[#fn:r488|488]]</sup> ). The few available observations are insufficient to detect trends. Where steep glaciers are frozen to bedrock, there is, however, ''medium evidence'' and ''high agreement'' from observations in the European Alps and from numerical simulations that failures of large parts of these glaciers were and will be facilitated in the future due to an increase in basal ice temperature (Fischer et al., 2013 <sup>[[#fn:r489|489]]</sup> ; Faillettaz et al., 2015 <sup>[[#fn:r490|490]]</sup> ; Gilbert et al., 2015 <sup>[[#fn:r491|491]]</sup> ) .

In some regions, glacier surges constitute a recurring hazard, due to widespread, quasi-periodic and substantial increases in glacier speed over a period of a few months to years, often accompanied by glacier advance (Harrison et al., 2015 <sup>[[#fn:r492|492]]</sup> ; Sevestre and Benn, 2015 <sup>[[#fn:r493|493]]</sup> ). In a number of cases, mostly in North America and High Mountain Asia (Bevington and Copland, 2014 <sup>[[#fn:r494|494]]</sup> ; Round et al., 2017 <sup>[[#fn:r495|495]]</sup> ; Steiner et al., 2018 <sup>[[#fn:r496|496]]</sup> ), surge-related glacier advances dammed rivers, causing major floods. In rare cases, glacier surges directly inundated agricultural land and damaged infrastructure (Shangguan et al., 2016 <sup>[[#fn:r497|497]]</sup> ). Sevestre and Benn (2015) <sup>[[#fn:r498|498]]</sup> suggest that surging operates within a climatic envelope of temperature and precipitation conditions, and that shifts in these conditions can modify surge frequencies and magnitudes. Some glaciers have reduced or stopped surge activity, or are projected to do so within decades, as a consequence of negative glacier mass balances (Eisen et al., 2001 <sup>[[#fn:r499|499]]</sup> ; Kienholz et al., 2017 <sup>[[#fn:r500|500]]</sup> ). For such cases, related hazards can also be expected to decrease. In contrast, intensive or increased surge activity (Hewitt, 2007 <sup>[[#fn:r501|501]]</sup> ; Gardelle et al., 2012 <sup>[[#fn:r502|502]]</sup> ; Yasuda and Furuya, 2015 <sup>[[#fn:r503|503]]</sup> ) occurred in a region on and around the Western Tibetan plateau which exhibited balanced or even positive glacier mass budgets in recent decades (Brun et al., 2017 <sup>[[#fn:r504|504]]</sup> ). Enhanced melt water production was suggested to be able to trigger or enhance surge-type instability, in particular for glaciers that contain ice both at the melting point and considerably below (Dunse et al., 2015 <sup>[[#fn:r505|505]]</sup> ; Yasuda and Furuya, 2015 <sup>[[#fn:r506|506]]</sup> ; Nuth et al., 2019 <sup>[[#fn:r507|507]]</sup> ).

A rare type of glacier instability with large volumes (in the order of 10-100 million m 3 ) and high mobility (up to 200–300 km/h) results from the complete collapse of large sections of low-angle valley glaciers and subsequent combined ice/rock/debris avalanches. The largest of such glacier collapses have been reported in the Caucasus Mountains in 2002 (Kolka Glacier, ~130 fatalities) (Huggel et al., 2005 <sup>[[#fn:r508|508]]</sup> ; Evans et al., 2009 <sup>[[#fn:r509|509]]</sup> ), and in the Aru Range in Tibet in 2016 (twin glacier collapses with 9 fatalities) (Kääb et al., 2018 <sup>[[#fn:r510|510]]</sup> ). Although there is no evidence that climate change has played a direct role in the 2002 event, changes in glacier mass balance, water input into the glaciers, and the frozen regime of the glacier beds were involved in the 2016 collapses and at least partly linked with climate change (Gilbert et al., 2018 <sup>[[#fn:r511|511]]</sup> ). Besides the 2016 Tibet cases, it is unknown if such massive and rare collapse-like glacier instabilities can be attributed to climate change.

<div id="section-2-3-2-1observed-and-projected-changes-block-2"></div>

<span id="snow-avalanches"></span>
===== 2.3.2.1.2 Snow avalanches =====

Snow avalanches can occur either spontaneously due to meteorological factors such as loading by snowfall or liquid water infiltration following, for example, surface melt or rain-on-snow, or can be triggered by the passage of people in avalanche terrain, the impact of falling ice or rocks, or by explosives used for avalanche control (Schweizer et al., 2003 <sup>[[#fn:r512|512]]</sup> ). There is no published evidence found that addresses the links between climate change and accidental avalanches triggered by recreationists or workers. Changes in snow cover characteristics are expected to induce changes in spontaneous avalanche activity including changes in friction and flow regime (Naaim et al., 2013 <sup>[[#fn:r513|513]]</sup> ; Steinkogler et al., 2014 <sup>[[#fn:r514|514]]</sup> ).

Ballesteros-Cánovas et al. (2018) <sup>[[#fn:r515|515]]</sup> reported increased avalanche activity in some slopes of the Western Indian Himalaya over the past decades related to increased frequency of wet-snow conditions. In the European Alps, avalanche numbers and runout distance have decreased where snow depth decreased and air temperature increased (Teich et al., 2012 <sup>[[#fn:r516|516]]</sup> ; Eckert et al., 2013 <sup>[[#fn:r517|517]]</sup> ). In the European Alps and Tatras mountains, over past decades, there has been a decrease in avalanche mass and run-out distance and a decrease in avalanches with a powder part; avalanche numbers decreased below 2,000 m a.s.l., and increased above (Eckert et al., 2013 <sup>[[#fn:r518|518]]</sup> ; Lavigne et al., 2015 <sup>[[#fn:r519|519]]</sup> ; Gadek et al., 2017 <sup>[[#fn:r520|520]]</sup> ). A positive trend in the proportion of avalanches involving wet snow in December through February was shown for the last decades (Pielmeier et al., 2013 <sup>[[#fn:r521|521]]</sup> ; Naaim et al., 2016 <sup>[[#fn:r522|522]]</sup> ). Land use and land cover changes also contributed to changes in avalanches (García-Hernández et al., 2017 <sup>[[#fn:r523|523]]</sup> ; Giacona et al., 2018 <sup>[[#fn:r524|524]]</sup> ). Correlations between avalanche activity and the El Niño-Southern Oscillation (ENSO) were identified from 1950–2011 in North and South America but there was no significant temporal trend reported for avalanche activity (McClung, 2013 <sup>[[#fn:r525|525]]</sup> ). Mostly inconclusive results were reported by Sinickas et al. (2015) <sup>[[#fn:r526|526]]</sup> and Bellaire et al. (2016) <sup>[[#fn:r527|527]]</sup> regarding the relationship between avalanche activity, climate change and disaster risk reduction activities in North America. In summary, in particular in Europe, there is ''medium confidence'' in an increase in avalanche activity involving wet snow, and a decrease in the size and run-out distance of snow avalanches over the past decades.

Future projections mostly indicate an overall decrease in snow depth and snow cover duration at lower elevation (Section 2.2.2), but the probability of occurrence of occasionally large snow precipitation events is projected to remain possible throughout most of the 21st century (Section 2.2.1). Castebrunet et al. (2014) <sup>[[#fn:r528|528]]</sup> estimated an overall 20 and 30% decrease of natural avalanche activity in the French Alps for the mid and end of the 21st century, respectively, under A1B scenario, compared to the reference period 1960–1990. Katsuyama et al. (2017) <sup>[[#fn:r529|529]]</sup> reached similar conclusions for Northern Japan, and Lazar and Williams (2008) <sup>[[#fn:r530|530]]</sup> for North America. Avalanches involving wet snow are projected to occur more frequently during the winter at all elevations due to surface melt or rain-on-snow (e.g., Castebrunet et al., 2014, for the French Alps), and the overall number and runout distance of snow avalanches is projected to decrease in regions and elevations experiencing significant reduction in snow cover (Mock et al., 2017 <sup>[[#fn:r531|531]]</sup> ). In summary, there is ''medium evidence'' and ''high agreement'' that observed changes in avalanches in mountain regions will be exacerbated in the future, with generally a decrease in hazard at lower elevation, and mixed changes at higher elevation (increase in avalanches involving wet snow, no clear direction of trend for overall avalanche activity).

<div id="section-2-3-2-1observed-and-projected-changes-block-3"></div>

<span id="figure-2.7"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.7'''

<span id="figure-2.7-anticipated-changes-in-high-mountain-hazards-under-climate-change-driven-by-changes-in-snow-cover-glaciers-and-permafrost-overlay-changes-in-the-exposure-and-vulnerability-of-individuals-communities-and-mountain-infrastructure."></span>
<!-- IMG CAPTION -->
'''Figure 2.7 | Anticipated changes in high mountain hazards under climate change, driven by changes in snow cover, glaciers and permafrost, overlay changes in the exposure and vulnerability of individuals, communities, and mountain infrastructure.'''
<!-- IMG FILE -->
[[File:209432aa6405ea2c6e267b9e97b74276 IPCC-SROCC-CH_2_7.jpg]]

Figure 2.7 | Anticipated changes in high mountain hazards under climate change, driven by changes in snow cover, glaciers and permafrost, overlay changes in the exposure and vulnerability of individuals, communities, and mountain infrastructure.

<!-- END IMG -->
<div id="section-2-3-2-1observed-and-projected-changes-block-4">
</div>

<span id="floods"></span>
===== 2.3.2.1.3 Floods =====

Glacier-related floods, including floods from lake outbursts (GLOFs), are documented for most glacierised mountain ranges and are among the most far-reaching glacier hazards. Past events affected areas tens to hundreds of kilometres downstream (Carrivick and Tweed, 2016 <sup>[[#fn:r532|532]]</sup> ). Retreating glaciers produced lakes at their fronts in many high mountain regions in recent decades (Frey et al., 2010 <sup>[[#fn:r533|533]]</sup> ; Gardelle et al., 2011 <sup>[[#fn:r534|534]]</sup> ; Loriaux and Casassa, 2013 <sup>[[#fn:r535|535]]</sup> ). Lake systems in High Mountain Asia also often developed on the surface of downwasting, low-slope glaciers where they coalesced from temporally variable supraglacial lakes (Benn et al., 2012 <sup>[[#fn:r536|536]]</sup> ; Narama et al., 2017 <sup>[[#fn:r537|537]]</sup> ). Corroborating SREX and AR5 findings, there is ''high confidence'' that current global glacier shrinkage caused new lakes to form and existing lakes to grow in most regions, for instance in South America, High mountain Asia and Europe (Loriaux and Casassa, 2013 <sup>[[#fn:r538|538]]</sup> ; Paul and Mölg, 2014 <sup>[[#fn:r539|539]]</sup> ; Zhang et al., 2015 <sup>[[#fn:r540|540]]</sup> ; Buckel et al., 2018 <sup>[[#fn:r541|541]]</sup> ). Exceptions occurred and are expected to occur in the future for few lakes where evaporation, runoff and reduced melt water influx in total led to a negative water balance (Sun et al., 2018a <sup>[[#fn:r542|542]]</sup> ). Also, advancing glaciers temporarily dammed rivers, lake sections, or fjords (Stearns et al., 2015 <sup>[[#fn:r543|543]]</sup> ), for instance through surging (Round et al., 2017 <sup>[[#fn:r544|544]]</sup> ), causing particularly large floods once the ice dams breached. Outbursts from water bodies in and under glaciers are able to cause floods similar to those from surface lakes but little is known about the processes involved and any trends under climate change. In some cases, the glacier thermal regime played a role so that climate driven changes in thermal regime are expected to alter the hazard potential, depending on local conditions (Gilbert et al., 2012 <sup>[[#fn:r545|545]]</sup> ). Another source of large water bodies under glaciers and subsequent floods has been subglacial volcanic activity (Section 2.3.2.1.4). There is also ''high confidence'' that the number and area of glacier lakes will continue to increase in most regions in the coming decades, and new lakes will develop closer to steep and potentially unstable mountain walls where lake outbursts can be more easily triggered by the impact of landslides (Frey et al., 2010 <sup>[[#fn:r546|546]]</sup> ; ICIMOD, 2011 <sup>[[#fn:r547|547]]</sup> ; Allen et al., 2016a <sup>[[#fn:r548|548]]</sup> ; Linsbauer et al., 2016 <sup>[[#fn:r549|549]]</sup> ; Colonia et al., 2017 <sup>[[#fn:r550|550]]</sup> ; Haeberli et al., 2017 <sup>[[#fn:r551|551]]</sup> ).

In contrast to the number and size of glacier lakes, trends in the number of glacier-related floods are not well known for recent decades (Carrivick and Tweed, 2016 <sup>[[#fn:r552|552]]</sup> ; Harrison et al., 2018 <sup>[[#fn:r553|553]]</sup> ), although a number of periods of increased and decreased flood activity have been documented for individual glaciers in North America and Greenland, spanning decades (Geertsema and Clague, 2005 <sup>[[#fn:r554|554]]</sup> ; Russell et al., 2011 <sup>[[#fn:r555|555]]</sup> ). A decrease in moraine-dammed glacier lake outburst floods in recent decades suggests a response of lake outburst activity being delayed by some decades with respect to glacier retreat (Harrison et al., 2018 <sup>[[#fn:r556|556]]</sup> ) but inventories might significantly underestimate the number of events (Veh et al., 2018 <sup>[[#fn:r557|557]]</sup> ). For the Himalaya, Veh et al. (2019) found no increase in the number of glacier lake outburst floods since the late 1980s. The degradation of permafrost and the melting of ice buried in lake dams have been shown to lower dam stability and contribute to outburst floods in many high mountain regions (Fujita et al., 2013 <sup>[[#fn:r558|558]]</sup> ; Erokhin et al., 2017 <sup>[[#fn:r559|559]]</sup> ; Narama et al., 2017 <sup>[[#fn:r560|560]]</sup> ).

Floods originating from the combination of rapidly melting snow and intense rainfall, referred to as rain-on-snow events, are some of the most damaging floods in mountain areas (Pomeroy et al., 2016 <sup>[[#fn:r561|561]]</sup> ; Il Jeong and Sushama, 2018). The hydrological response of a catchment to a rain-on-snow event depends on the characteristics of the precipitation event, but also on turbulent fluxes driven by wind and humidity, which typically provide most of the melting energy during such events (Pomeroy et al., 2016 <sup>[[#fn:r562|562]]</sup> ), and the state of the snowpack, in particular the liquid water content (Würzer et al., 2016 <sup>[[#fn:r563|563]]</sup> ). An increase in the occurrence of rain-on-snow events in high-elevation zones, and a decrease at the lowest elevations have been reported (Western USA, 1949–2003, McCabe et al. (2007); Oregon, 1986–2010, Surfleet and Tullos (2013) <sup>[[#fn:r564|564]]</sup> ; Switzerland, 1972–2016, Moran-Tejéda et al. (2016), central Europe, 1950–2010, Freudiger et al. (2014) <sup>[[#fn:r565|565]]</sup> . These trends are consistent with studies carried out at the scale of the Northern Hemisphere (Putkonen and Roe, 2003 <sup>[[#fn:r566|566]]</sup> ; Ye et al., 2008 <sup>[[#fn:r567|567]]</sup> ; Cohen et al., 2015 <sup>[[#fn:r568|568]]</sup> ). There are no studies found on this topic in Africa and South America. In summary, evidence since AR5 suggests that rain-on-snow events have increased over the last decades at high elevations, particularly during transition periods from autumn to winter and winter to spring ( ''medium confidence'' ). The occurrence of rain-on-snow events has decreased over the last decade in low-elevation or low-latitude areas due to a decreasing duration of the snowpack, except for the coldest months of the year ( ''medium confidence'' ).

Il Jeong and Sushama (2018) projected an increase in rain-on-snow events in winter and a decrease in spring, for the period 2041–2070 (RCP4.5 and RCP8.5) in North America, corroborated by Musselman et al. (2018). Their frequency in the Swiss Alps is projected to increase at elevations higher than 2,000 m a.s.l. (SRES A1B, 2025, 2055, and 2085) ( Beniston and Stoffel, 2016 <sup>[[#fn:r569|569]]</sup> ) . This study showed that the number of rain-on-snow events may increase by 50%, with a regional temperature increase of 2°C to 4°C, and decrease with a temperature rise exceeding 4°C. In Alaska, an overall increase of rain-on-snow events is projected, however with a projected decline in the southwestern/southern region (Bieniek et al., 2018 <sup>[[#fn:r570|570]]</sup> ). In summary, evidence since AR5 suggests that the frequency of rain-on-snow events is projected to increase and occur earlier in spring and later in autumn at higher elevation and to decrease at lower elevation ( ''high confidence'' ).

<div id="section-2-3-2-1observed-and-projected-changes-block-5"></div>

<span id="combined-hazards-and-cascading-events"></span>
===== 2.3.2.1.4 Combined hazards and cascading events =====

The largest mountain disasters in terms of reach, damage and lives lost that involve ice, snow and permafrost occurred through a combination or chain of processes. New evidence since SREX and AR5 have these findings (Anacona et al., 2015a <sup>[[#fn:r571|571]]</sup> ; Evans and Delaney, 2015) <sup>[[#fn:r572|572]]</sup> . Some process chains occur frequently, while others are rare, specific to local circumstances and difficult to anticipate. Glacier lake outbursts were in many mountain regions and over recent decades documented to have been triggered by impact waves from snow-, ice- or rock-avalanches, landslides, iceberg calving events, or by temporary blockage of surface or subsurface drainage channels (Benn et al., 2012 <sup>[[#fn:r573|573]]</sup> ; Narama et al., 2017 <sup>[[#fn:r574|574]]</sup> ). Rock-slope instability and catastrophic failure along fjords caused tsunamis (Hermanns et al., 2014 <sup>[[#fn:r575|575]]</sup> ; Roberts et al., 2014 <sup>[[#fn:r576|576]]</sup> ). For instance, a landslide generated wave in 2015 at Taan Fjord, Alaska, ran up 193 m on the opposite slope and then travelled more than 20 km down the fjord (Higman et al., 2018 <sup>[[#fn:r577|577]]</sup> ). Earthquakes have been a starting point for different types of cascading events, for instance by causing snow-, ice- or rock-avalanches, and landslides (van der Woerd et al., 2004 <sup>[[#fn:r578|578]]</sup> ; Podolskiy et al., 2010 <sup>[[#fn:r579|579]]</sup> ; Cook and Butz, 2013 <sup>[[#fn:r580|580]]</sup> ; Sæmundsson et al., 2018 <sup>[[#fn:r581|581]]</sup> ). Glaciers and their moraines, including morainic lake dams, seem however, not particularly prone to earthquake triggered failure (Kargel et al., 2016 <sup>[[#fn:r582|582]]</sup> ).

Landslides and rock avalanches in glacier environments were often documented to entrain snow and ice that fluidise, and incorporate additional loose glacial sediments or water bodies, thereby multiplying their mobility, volume and reach (Schneider et al., 2011 <sup>[[#fn:r583|583]]</sup> ; Evans and Delaney, 2015 <sup>[[#fn:r584|584]]</sup> ). Rock avalanches onto glaciers triggered glacier advances in recent decades, for instance in North America, New Zealand and Europe, mainly through reducing surface melt (Deline, 2009 <sup>[[#fn:r585|585]]</sup> ; Reznichenko et al., 2011 <sup>[[#fn:r586|586]]</sup> ; Menounos et al., 2013 <sup>[[#fn:r587|587]]</sup> ). In glacier covered frozen rock walls, particularly complex thermal, mechanical, hydraulic and hydrologic interactions between steep glaciers, frozen rock and its ice content, and unfrozen rock sections lead to combined rock/ice instabilities that are difficult to observe and anticipate (Harris et al., 2009 <sup>[[#fn:r588|588]]</sup> ; Fischer et al., 2013 <sup>[[#fn:r589|589]]</sup> ; Ravanel et al., 2017 <sup>[[#fn:r590|590]]</sup> ). There is ''limited evidence'' of observed direct event chains to project future trends. However, from the observed and projected degradation of permafrost, shrinkage of glaciers and increase in glacier lakes it is reasonable to assume that event chains involving these could increase in frequency or magnitude, and that accordingly hazard zones could expand.

Volcanoes covered by snow and ice often produce substantial melt water during eruptions. This typically results in floods and/or lahars (mixtures of melt water and volcanic debris) which can be exceptionally violent and cause large-scale loss of life and destruction to infrastructure (Barr et al., 2018 <sup>[[#fn:r591|591]]</sup> ). The most devastating example from recent history occurred in 1985, when the medium-sized eruption of Nevado del Ruiz volcano, Colombia, produced lahars that killed more than 23,000 people some 70 km downstream (Pierson et al., 1990 <sup>[[#fn:r592|592]]</sup> ). Hazards associated with ice and snow-clad volcanoes have been reported mostly from the Cordilleras of the Americas, but also from the Aleutian arc (USA), Mexico, Kamchatka (Russia), Japan, New Zealand and Iceland (Seynova et al., 2017 <sup>[[#fn:r593|593]]</sup> ). In particular, under Icelandic glaciers, volcanic activity and eruptions melted large amounts of ice and caused especially large floods if water accumulated underneath the glacier (Björnsson, 2003 <sup>[[#fn:r594|594]]</sup> ; Seneviratne et al., 2012 <sup>[[#fn:r595|595]]</sup> ). There is ''medium confidence'' that the overall hazard related to floods and lahars from ice- and snow-clad volcanoes will gradually diminish over years-to-decades as glaciers and seasonal snow cover continue to decrease under climate change (Aguilera et al., 2004 <sup>[[#fn:r596|596]]</sup> ; Barr et al., 2018 <sup>[[#fn:r597|597]]</sup> ). On the other hand, shrinkage of glaciers may uncover steep slopes of unconsolidated volcanic sediments, thus decreasing in the future the resistance of these volcano flanks to heavy rain fall and increasing the hazard from related debris flows (Vallance, 2005 <sup>[[#fn:r598|598]]</sup> ). In summary, future changes in snow and ice are expected to modify the impacts of volcanic activity of snow- and ice-clad volcanoes ( ''high confidence)'' although in complex and locally variable ways and at a variety of time scales (Barr et al., 2018 <sup>[[#fn:r599|599]]</sup> ; Swindles et al., 2018 <sup>[[#fn:r600|600]]</sup> ).

<div id="section-2-3-2-2exposure-vulnerability-and-impacts"></div>

<span id="exposure-vulnerability-and-impacts"></span>