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

== 2.3 Mountain Social-Ecological Systems: Impacts, Risks and Human Responses ==

<span id="water-resources"></span>
=== 2.3.1 Water Resources ===

<div id="section-2-3-1water-resources-block-1"></div>

The mountain cryosphere is an important source of freshwater in the mountains themselves and in downstream regions. The runoff per unit area generated in mountains is on average approximately twice as high as in lowlands (Viviroli et al., 2011 <sup>[[#fn:r237|237]]</sup> ) making mountains a significant source of fresh water in sustaining ecosystem and supporting livelihoods in and far beyond the mountain ranges themselves. The presence of snow, glaciers, and permafrost generally exert a strong control on the amount, timing and biogeochemical properties of runoff (FAQ 2.1). Changes to the cryosphere due to climate change can alter freshwater availability with direct consequences for human populations and ecosystems

<div id="section-2-3-1-1changes-in-river-runoff"></div>

<span id="changes-in-river-runoff"></span>
==== 2.3.1.1 Changes in River Runoff ====

<div id="section-2-3-1-1changes-in-river-runoff-block-1"></div>

AR5 reported increased winter flows and a shift in timing towards earlier spring snowmelt runoff peaks during previous decades ( ''robust evidence, high agreement'' ). In glacier-fed river basins, it was projected that melt water yields from glaciers will increase for decades in many regions but then decline ( ''very high confidence'' ). These findings have been further supported and refined by a wealth of new studies since AR5.

Recent studies indicate considerable changes in the seasonality of runoff in snow and glacier dominated river basins ( ''very high confidence'' ; Table SM2.9). Several studies have reported an increase in average winter runoff over the past decades, for example in Western Canada (Moyer et al., 2016 <sup>[[#fn:r238|238]]</sup> ), the European Alps (Bocchiola, 2014 <sup>[[#fn:r239|239]]</sup> ; Bard et al., 2015 <sup>[[#fn:r240|240]]</sup> ) and Norway (Fleming and Dahlke, 2014 <sup>[[#fn:r241|241]]</sup> ), due to more precipitation falling as rain under warmer conditions. Summer runoff has been observed to decrease in basins, for example in Western Canada (Brahney et al., 2017 <sup>[[#fn:r242|242]]</sup> ) and the European Alps (Bocchiola, 2014 <sup>[[#fn:r243|243]]</sup> ), but to increase in several basins in High Mountain Asia (Mukhopadhyay and Khan, 2014 <sup>[[#fn:r244|244]]</sup> ; Duethmann et al., 2015 <sup>[[#fn:r245|245]]</sup> ; Reggiani and Rientjes, 2015 <sup>[[#fn:r246|246]]</sup> ; Engelhardt et al., 2017 <sup>[[#fn:r247|247]]</sup> ). Both increases, for example, in Alaska (Beamer et al., 2016 <sup>[[#fn:r248|248]]</sup> ) and the Tien Shan (Wang et al., 2015 <sup>[[#fn:r249|249]]</sup> ; Chen et al., 2016 <sup>[[#fn:r250|250]]</sup> ), and decreases, for example, in Western Canada (Brahney et al., 2017 <sup>[[#fn:r251|251]]</sup> ) have also been found for average annual runoff. In Western Austria, Kormann et al. (2015) detected an increase in annual flow at high elevations and a decrease at low elevations between 1980–2010.

These contrasting trends for summer and annual runoff often result from spatially variable changes in the contribution of glacier and snow melt. As glaciers shrink, annual glacier runoff typically first increases, until a turning point, often called ‘peak water’ is reached, upon which runoff declines (FAQ 2.1). There is ''robust evidence'' and ''high agreement'' that peak water in glacier-fed rivers has already passed with annual runoff declining especially in mountain regions with predominantly smaller glaciers, for example, in the tropical Andes (Frans et al., 2015 <sup>[[#fn:r252|252]]</sup> ; Polk et al., 2017 <sup>[[#fn:r253|253]]</sup> ), Western Canada (Fleming and Dahlke, 2014 <sup>[[#fn:r254|254]]</sup> ; Brahney et al., 2017 <sup>[[#fn:r255|255]]</sup> ) and the Swiss Alps (Huss and Fischer, 2016 <sup>[[#fn:r256|256]]</sup> ). A global modelling study (Huss and Hock, 2018 <sup>[[#fn:r257|257]]</sup> ) suggests that peak water has been reached before 2019 for 82–95% of the glacier area in the tropical Andes, 40–49% in Western Canada and USA, and 55–67% in Central Europe (including European Alps and Pyrenees) and the Caucasus (Figure 2.6).

Projections indicate a continued increase in winter runoff in many snow and/or glacier-fed rivers over the 21st century ( ''high confidence'' ) regardless of the climate scenario, for example, in North America (Schnorbus et al., 2014 <sup>[[#fn:r258|258]]</sup> ; Sultana and Choi, 2018 <sup>[[#fn:r259|259]]</sup> ), the European Alps (Addor et al., 2014 <sup>[[#fn:r260|260]]</sup> ; Bosshard et al., 2014 <sup>[[#fn:r261|261]]</sup> ), Scotland (Capell et al., 2014 <sup>[[#fn:r262|262]]</sup> ) and High Mountain Asia (Kriegel et al., 2013 <sup>[[#fn:r263|263]]</sup> ) due to increased winter snowmelt and more precipitation falling as rain in addition to increases in precipitation in some basins (Table SM2.9). There is ''robust evidence'' ( ''high agreement'' ) that summer runoff will decline over the 21st century in many basins for all emission scenarios, for example, in Western Canada and USA (Shrestha et al., 2017 <sup>[[#fn:r264|264]]</sup> ), the European Alps (Jenicek et al., 2018 <sup>[[#fn:r265|265]]</sup> ), High Mountain Asia (Prasch et al., 2013 <sup>[[#fn:r266|266]]</sup> ; Engelhardt et al., 2017 <sup>[[#fn:r267|267]]</sup> ) and the tropical Andes (Baraer et al., 2012 <sup>[[#fn:r268|268]]</sup> ), due to less snowfall and decreases in glacier melt after peak water. A global-scale projection suggests that decline in glacier runoff by 2100 (RCP8.5) may reduce basin runoff by 10% or more in at least one month of the melt season in several large river basins, especially in High Mountain Asia during dry seasons, despite glacier cover of less than a few percent (Huss and Hock, 2018 <sup>[[#fn:r269|269]]</sup> ).

Projected changes in annual runoff in glacier dominated basins are complex including increases and decreases over the 21st century for all scenarios depending on the time period and the timing of peak water ( ''high confidence'' ) (Figure 2.6). Local and regional-scale projections in High Mountain Asia, the European Alps, and Western Canada and USA suggest that peak water will generally be reached before or around the middle of the century. These finding are consistent with results from global-scale modelling of glacier runoff (Bliss et al., 2014 <sup>[[#fn:r270|270]]</sup> ; Huss and Hock, 2018 <sup>[[#fn:r271|271]]</sup> ) indicating generally earlier peak water in regions with little ice cover and smaller glaciers (e.g., Low Latitudes, European Alps and Pyrenees, and the Caucasus) and later peak water in regions with extensive ice cover and large glaciers (e.g., Alaska, Southern Andes). In some regions (e.g., Iceland) peak water from most glacier area is projected to occur earlier for RCP2.6 than RCP8.5, caused by decreasing glacier runoff as glaciers find a new equilibrium. In contrast melt-driven glacier runoff continues to rise for the higher emission scenario. There is ''very high confidence'' that spring peak runoff in many snow-dominated basins around the world will occur earlier in the year, up to several weeks, by the end of the century caused by earlier snowmelt (e.g., Coppola et al., 2014; Bard et al., 2015 <sup>[[#fn:r272|272]]</sup> ; Yucel et al., 2015 <sup>[[#fn:r273|273]]</sup> ; Islam et al., 2017 <sup>[[#fn:r274|274]]</sup> ; Sultana and Choi, 2018 <sup>[[#fn:r275|275]]</sup> ).

In addition to changes in ice and snow melt, changes in other variables such as precipitation and evapotranspiration due to atmospheric warming or vegetation change affect runoff amounts and timing (e.g., Bocchiola, 2014; Lutz et al., 2016 <sup>[[#fn:r276|276]]</sup> ). Changes in melt water from ice and snow often dominates the runoff response to climate change at higher elevations, while changes in precipitation and evapotranspiration become increasingly important at lower elevations (Kormann et al., 2015 <sup>[[#fn:r277|277]]</sup> ). Permafrost thaw may affect runoff by releasing water from ground ice (Jones et al., 2018 <sup>[[#fn:r278|278]]</sup> ) and indirectly by changing hydrological pathways or ground water recharge as permafrost degrades (Lamontagne-Hallé et al., 2018 <sup>[[#fn:r279|279]]</sup> ). The relative importance of runoff from thawing permafrost compared to runoff from melting glaciers is expected to be greatest in arid areas where permafrost tends to be more abundant (Gruber et al., 2017 <sup>[[#fn:r280|280]]</sup> ). Because glaciers react more rapidly to climate change than permafrost, runoff in some mountain landscapes may become increasingly affected by permafrost thaw in the future (Jones et al., 2018 <sup>[[#fn:r281|281]]</sup> ).

In summary, there is ''very high confidence'' that glacier and snow cover decline have affected and will continue to change the amounts and seasonality of river runoff in many snow-dominated and/or glacier-fed river basins. The average winter runoff is expected to increase ( ''high confidence'' ), and spring peak maxima will occur earlier ( ''very'' ''high confidence'' ). Although observed and projected trends in annual runoff vary substantially among regions and can even be opposite in sign, there is ''high confidence'' that average annual runoff from glaciers will have reached a peak, with declining runoff thereafter, at the latest by the end of the 21st century in all regions regardless emission scenario. The projected changes in runoff are expected to affect downstream water management, related hazards and ecosystems (Section 2.3.2, 2.3.3).

<div id="section-2-3-1-1changes-in-river-runoff-block-2"></div>

<span id="figure-2.6"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 2.6'''

<span id="figure-2.6-timing-of-peak-water-from-glaciers-in-different-regions-figure-2.1-under-two-emission-scenarios-for-representative-concentration-pathways-rcp2.6-and-rcp8.5.-peak-water-refers-to-the-year-when-annual-runoff-from-the-initially-glacier-covered-area-will-start-to-decrease-due-to-glacier-shrinkage-after-a-period-of-melt-induced-increase.-the"></span>
<!-- IMG CAPTION -->
'''Figure 2.6 | Timing of peak water from glaciers in different regions (Figure 2.1) under two emission scenarios for Representative Concentration Pathways RCP2.6 and RCP8.5. Peak water refers to the year when annual runoff from the initially glacier-covered area will start to decrease due to glacier shrinkage after a period of melt induced increase. The […]'''
<!-- IMG FILE -->
[[File:6198ac49c2ac634a93bee8a7d1491d5a IPCC-SROCC-CH_2_6.jpg]]

Figure 2.6 | Timing of peak water from glaciers in different regions (Figure 2.1) under two emission scenarios for Representative Concentration Pathways RCP2.6 and RCP8.5. Peak water refers to the year when annual runoff from the initially glacier-covered area will start to decrease due to glacier shrinkage after a period of melt induced increase. The bars are based on Huss and Hock (2018) who used a global glacier model to compute the runoff of all individual glaciers in a region until year 2100 based on 14 General Circulation Models (GCMs). Depicted is the area of all glaciers that fall into the same 10-year peak water interval expressed as a percentage of each region’s total glacier area, i.e., all bars for the same RCP sum up to 100% glacier area. Shadings of the bars distinguish different glacier sizes indicating a tendency for peak water to occur later for larger glaciers. Circles/diamonds mark timing of peak water from individual case studies based on observations or modelling (Table SM2.10). Circles refer to results from individual glaciers regardless of size or a collection of glaciers covering 150 km2 glacier coverage. Case studies based on observations or scenarios other than RCP2.6 and RCP8.5 are shown in both the left and right set of panels.

<!-- END IMG -->
<div id="section-2-3-1-2water-quality">
</div>

<span id="water-quality"></span>
==== 2.3.1.2 Water Quality ====

<div id="section-2-3-1-2water-quality-block-1"></div>

Glacier decline can influence water quality by accelerating the release of stored anthropogenic legacy pollutants, with impacts to downstream ecosystem services. These legacy pollutants notably include persistent organic pollutants (POPs), particularly polychlorinated biphenyls (PCBs) and dichlorodiphenyl-trichloroethane (DDT), polycyclic aromatic hydrocarbons, and heavy metals (Hodson, 2014 <sup>[[#fn:r282|282]]</sup> ) and are associated with the deposition and release of black carbon. There is ''limited evidence'' that some of these pollutants found in surface waters in the Gangetic Plain during the dry season originate from Himalayan glaciers (Sharma et al., 2015 <sup>[[#fn:r283|283]]</sup> ), and glaciers in the European Alps store the largest known quantity of POPs in the Northern Hemisphere (Milner et al., 2017 <sup>[[#fn:r284|284]]</sup> ). Although their use has declined or ceased worldwide, PCBs have been detected in runoff from glacier melt due to the lag time of release from glaciers (Li et al., 2017 <sup>[[#fn:r285|285]]</sup> ). Glaciers also represent the most unstable stores of DDT in European and other mountain areas flanking large urban centres and glacier derived DDT is still accumulating in lake sediments downstream from glaciers (Bogdal et al., 2010 <sup>[[#fn:r286|286]]</sup> ). However, bioflocculation (the aggregation of dispersed organic particles by the action of organisms) can increase the residence time of these contaminants stored in glaciers thereby reducing their overall toxicity to freshwater ecosystems (Langford et al., 2010 <sup>[[#fn:r287|287]]</sup> ). Overall the effect on freshwater ecosystems of these contaminants is estimated to be low ( ''medium confidence'' ) (Milner et al., 2017 <sup>[[#fn:r288|288]]</sup> ).

Of the heavy metals, mercury is of particular concern and an estimated 2.5 tonnes has been released by glaciers to downstream ecosystems across the Tibetan Plateau over the last 40 years (Zhang et al., 2012 <sup>[[#fn:r289|289]]</sup> ). Mercury in glacial silt, originating from grinding of rocks as the glacier flows over them, can be as large or larger than the mercury flux from melting ice due to anthropogenic sources deposited on the glacier (Zdanowicz et al., 2013 <sup>[[#fn:r290|290]]</sup> ). Both glacier erosion and atmospheric deposition contributed to the high rates of total mercury export found in a glacierised watershed in coastal Alaska (Vermilyea et al., 2017 <sup>[[#fn:r291|291]]</sup> ) and mercury output is predicted to increase in glacierised mountain catchments (Sun et al., 2017 <sup>[[#fn:r292|292]]</sup> ; Sun et al., 2018b <sup>[[#fn:r293|293]]</sup> ) ( ''medium confidence'' ). However, a key issue is how much of this glacier-derived mercury, largely in the particulate form, is converted to toxic methyl mercury downstream. Methyl mercury can be incorporated into aquatic food webs in glacier streams (Nagorski et al., 2014 <sup>[[#fn:r294|294]]</sup> ) and bio-magnify up the food chain (Lavoie et al., 2013 <sup>[[#fn:r295|295]]</sup> ). Water originating from rock glaciers can also contribute other heavy metals that exceed guideline values for drinking water quality (Thies et al., 2013 <sup>[[#fn:r296|296]]</sup> ). In addition, permafrost degradation can enhance the release of other trace elements (e.g., aluminium, manganese and nickel) (Colombo et al., 2018 <sup>[[#fn:r297|297]]</sup> ). Indeed, projections indicate that all scenarios of future climate change will enhance the mobilisation of metals in metamorphic mountain catchments (Zaharescu et al., 2016 <sup>[[#fn:r298|298]]</sup> ). The release of toxic contaminants, particularly where glacial melt waters are used for irrigation and drinking water in the Himalayas and the Andes, is potentially harmful to human health both now and in the future (Hodson, 2014 <sup>[[#fn:r299|299]]</sup> ) ( ''medium confidence)'' .

Soluble reactive phosphorus concentrations in rivers downstream of glaciers are predicted to decrease with declining glacier coverage (Hood et al., 2009 <sup>[[#fn:r300|300]]</sup> ) as a large percentage is associated with glacier-derived suspended sediment (Hawkings et al., 2016 <sup>[[#fn:r301|301]]</sup> ). In contrast, dissolved organic carbon (DOC), dissolved inorganic nitrogen and dissolved organic nitrogen concentrations in pro-glacial rivers is projected to increase this century due to glacier shrinkage (Hood et al., 2015 <sup>[[#fn:r302|302]]</sup> ; Milner et al., 2017 <sup>[[#fn:r303|303]]</sup> ) ( ''robust evidence, medium agreement)'' . Globally, mountain glaciers are estimated to release about 0.8 Tera g yr -1 (Li et al., 2018 <sup>[[#fn:r304|304]]</sup> ) of highly bioavailable DOC that may be incorporated into downstream food webs (Fellman et al., 2015 <sup>[[#fn:r305|305]]</sup> ; Hood et al., 2015 <sup>[[#fn:r306|306]]</sup> ). Loss rates of DOC from glaciers in the high mountains of the Tibetan Plateau were estimated to be ∼ 0.19 Tera g C yr -1 , (Li et al., 2018 <sup>[[#fn:r307|307]]</sup> ) higher than other regions suggesting that DOC is released more efficiently from Asian mountain glaciers (Liu et al., 2016 <sup>[[#fn:r308|308]]</sup> ). Glacier DOC losses are expected to accelerate as they shrink, leading to a cumulative annual loss of roughly 15 Tera g C yr -1  of glacial DOC by 2050 from melting glaciers and ice sheets (Hood et al., 2015 <sup>[[#fn:r309|309]]</sup> ). Permafrost degradation is also a major and increasing source of bioavailable DOC (Abbott et al., 2014 <sup>[[#fn:r310|310]]</sup> ; Aiken et al., 2014 <sup>[[#fn:r311|311]]</sup> ). Major ions calcium, magnesium, sulphate and nitrate (Colombo et al., 2018 <sup>[[#fn:r312|312]]</sup> ) are also released by permafrost degradation as well as acid drainage leaching into alpine lakes (Ilyashuk et al., 2018 <sup>[[#fn:r313|313]]</sup> ).

Increasing water temperature has been reported in some high mountain streams (e.g., Groll et al., 2015; Isaak et al., 2016 <sup>[[#fn:r314|314]]</sup> ) due to decreases in glacial runoff, producing changes in water quality and species richness (Section 2.3.3). In contrast, water temperature in regions with extensive glacier cover are expected to show a transient decline, due to an enhanced cooling effect from increased glacial melt water (Fellman et al., 2014 <sup>[[#fn:r315|315]]</sup> ).

In summary, changes in the mountain cryosphere will cause significant shifts in downstream nutrients (DOC, nitrogen, phosphorus) and influence water quality through increases in heavy metals, particularly mercury, and other legacy contaminants ( ''medium evidence, high agreement'' ) posing a potential threat to human health. These threats are more focused where glaciers are subject to substantial pollutant loads such as High Mountain Asia and Europe, rather than areas like Alaska and Canada.

<div id="section-2-3-1-3key-impacts-and-vulnerability"></div>

<span id="key-impacts-and-vulnerability"></span>
==== 2.3.1.3 Key Impacts and Vulnerability ====

<div id="section-2-3-1-3key-impacts-and-vulnerability-block-1"></div>

<span id="hydropower"></span>
===== 2.3.1.3.1 Hydropower =====

Hydropower comprises about 16% of electricity generation globally but close to 100%, in many mountainous countries (Hamududu and Killingtveit, 2012 <sup>[[#fn:r316|316]]</sup> ; IHA, 2018). It represents a significant source of revenue for mountainous regions (Gaudard et al., 2016 <sup>[[#fn:r317|317]]</sup> ). Due to the dependence on water resources as key input, hydropower operations are expected to be affected by changes in runoff from glaciers and snow cover (Section 2.3.1.1, FAQ 2.1). Both increases and decreases in annual and/or seasonal water input to hydropower facilities have been recorded in several high mountain regions, for example, in Switzerland (Hänggi and Weingartner, 2012 <sup>[[#fn:r318|318]]</sup> ; Schaefli et al., 2019 <sup>[[#fn:r319|319]]</sup> ), Canada (Jost et al., 2012 <sup>[[#fn:r320|320]]</sup> ; Jost and Weber, 2013 <sup>[[#fn:r321|321]]</sup> ), Iceland (Einarsson and Jónsson, 2010 <sup>[[#fn:r322|322]]</sup> ) and High Mountain Asia (Ali et al., 2018 <sup>[[#fn:r323|323]]</sup> ). However, there is only ''limited evidence'' ( ''medium agreement'' ) that changes in runoff have led to changes in hydropower plant operation. For example, in Iceland, the National Power Company observed in 2005 that flows into their energy system were greater than historical flows. By incorporating the most recent runoff data into strategies for reservoir management it was possible to increase production capacity (Braun and Fournier, 2016 <sup>[[#fn:r324|324]]</sup> ).

There is ''robust evidence (medium agreement'' ) that water input to hydropower facilities will change in the future due to cryosphere-related impacts on runoff (Section 2.3.1.1). For example, in the Skagit river basin in British Columbia and Northern Washington (Lee et al., 2016 <sup>[[#fn:r325|325]]</sup> ) and in California (Madani and Lund, 2010 <sup>[[#fn:r326|326]]</sup> ) projections (SRES A1B) show more runoff in winter and less in summer. In India, snow and glacier runoff to hydropower plants is projected to decline in several basins (Ali et al., 2018 <sup>[[#fn:r327|327]]</sup> ). In some cases, catchments that are close together are projected to evolve in contrasting directions in terms of runoff, for example in the European Alps (Gaudard et al., 2013 <sup>[[#fn:r328|328]]</sup> ; Gaudard et al., 2014 <sup>[[#fn:r329|329]]</sup> ). Increased runoff due to changes in the cryosphere will increase the risk of overflows (non-productive discharge), particularly during winter and spring melt, with the greatest impacts on run-of-river power plants (e.g., in Canada; Minville et al., 2010 <sup>[[#fn:r330|330]]</sup> ; Warren and Lemmen, 2014 <sup>[[#fn:r331|331]]</sup> ) ( ''medium confidence'' ).

There is ''medium evidence'' ( ''high agreement'' ) that changes in glacier- and moraine-dammed lakes, and changes in sediment supply will affect hydropower generation (Colonia et al., 2017 <sup>[[#fn:r332|332]]</sup> ; Hauer et al., 2018 <sup>[[#fn:r333|333]]</sup> ). Many glacier lakes have increased in volume, and can damage hydropower infrastructure when they empty suddenly (Engeset et al., 2005 <sup>[[#fn:r334|334]]</sup> ; Jackson and Ragulina, 2014 <sup>[[#fn:r335|335]]</sup> ; Carrivick and Tweed, 2016 <sup>[[#fn:r336|336]]</sup> ) (Section 2.3.2). If large enough, hydropower reservoirs can reduce the downstream negative impacts of changes in the cryosphere by storing and providing freshwater during hot, dry periods or by alleviating the effects of glacier floods (Jackson and Ragulina, 2014 <sup>[[#fn:r337|337]]</sup> ; Colonia et al., 2017 <sup>[[#fn:r338|338]]</sup> ). In mountain rivers, sediment volume and type depend on connectivity between hillslopes and the valley floor (Carrivick et al., 2013 <sup>[[#fn:r339|339]]</sup> ), glacier activity (Lane et al., 2017 <sup>[[#fn:r340|340]]</sup> ) and on water runoff regime feedbacks with river channel dynamics (Schmidt and Morche, 2006 <sup>[[#fn:r341|341]]</sup> ). An increase in suspended sediment loading under current reservoir operating policies is projected for some hydropower facilities, for example, in British Columbia and Northern Washington (Lee et al., 2016 <sup>[[#fn:r342|342]]</sup> ).

Only a few studies have addressed the economic effects on hydropower due directly to changes in the cryosphere. For example in Peru, Vergara et al. (2007) studied the effect of both reduced glacier runoff and runoff with no glacier input once the glaciers have completely melted for the Cañón del Pato hydropower plant in Peru, and found an economic cost of between 5 – 20 million USD yr -1 , with the lower figure for the cost of energy paid to the producer and the higher figure the society cost. Costs calculated for all of Peru, where ~80% of electricity comes from hydropower range from 60 – 212 million USD yr -1 . If the cost of rationing energy is considered, the national cost is estimated as 1,500 million USD yr -1 .

Other factors than changes in the cryosphere, such as market policies and regulation, may have greater significance for socioeconomic development of hydropower in the future (Section 2.3.1.4, Gaudard et al., 2016). Hence, despite the efforts of hydropower agencies and regulatory bodies to quantify changes or to develop possible adaptation strategies (IHA, 2018), only a few organisations are incorporating current knowledge of climate change into their investment planning. The World Bank uses a decision tree approach to identify potential vulnerabilities in a hydropower project incurred from key uncertain factors and their combinations (Bonzanigo et al., 2015 <sup>[[#fn:r343|343]]</sup> ).

<div id="section-2-3-1-3key-impacts-and-vulnerability-block-2"></div>

<span id="agriculture"></span>
===== 2.3.1.3.2 Agriculture =====

High mountains have supported agricultural livelihoods for centuries. Rural communities are dependent on adequate levels of soil moisture at planting time, derived in part in many cases from irrigation water which includes glacier and snowmelt water; as a result, they are exposed to risk which stems from cryosphere changes ( ''high confidence'' ) (Figure 2.8). The relative poverty of many mountain communities contributes to their vulnerability to the impacts of these cryosphere changes (McDowell et al., 2014 <sup>[[#fn:r344|344]]</sup> ; Carey et al., 2017 <sup>[[#fn:r345|345]]</sup> ; Rasul and Molden, 2019 <sup>[[#fn:r346|346]]</sup> ) ''(medium evidence, high agreement).'' Glacier and snowmelt water contribute irrigation water to adjacent lowlands as well. Pastoralism, an important livelihood strategy in mountain regions, is also impacted by cryosphere changes, but described in Section 2.3.7.

There is ''medium evidence (medium agreement)'' that reduction in streamflow due to glacier retreat or reduced snow cover has led to reduced water availability for irrigation of crops and declining agricultural yields in several mountain areas (Table SM2.11), for example in the tropical Andes (e.g., Bury et al., 2011) and High Mountain Asia (e.g., Nüsser and Schmidt, 2017). In the Southern Andes, increased streamflow in the Elqui River in Chile, due to glacier retreat or changing snow cover, has led to increased water availability for irrigation and increased agricultural yields (Young et al., 2010 <sup>[[#fn:r347|347]]</sup> ).

In addition to the effects on agriculture of changing availability of irrigation water, reductions in snow cover can also impact agriculture through its direct effects on soil moisture, as reported for Nepal, where lesser snow cover has led to the drying of soils and lower yields of potatoes and fodder (Smadja et al., 2015 <sup>[[#fn:r348|348]]</sup> ). Agriculture in high mountain areas is sensitive to other climatic drivers as well. Rising air temperatures increase crop evapotranspiration, thus increasing water demand for crop production to maintain optimal yield (Beniston and Stoffel, 2014 <sup>[[#fn:r349|349]]</sup> ). They are also associated with upslope movement of cropping zones, which favours some farmers in high mountain areas, who are increasingly able to cultivate new crops, such as onions, garlic and apples in Nepal (Huntington et al., 2017 <sup>[[#fn:r350|350]]</sup> ; Hussain et al., 2018 <sup>[[#fn:r351|351]]</sup> ), and maize in Ecuador (Skarbø and VanderMolen, 2014 <sup>[[#fn:r352|352]]</sup> ). Dry spells and unseasonal frosts have also impacted agriculture in Peru (Bury et al., 2011 <sup>[[#fn:r353|353]]</sup> ).

Adaptation activities in mountain agriculture related at least partially to cryospheric changes are detailed in Table SM2.12 and their geographic spread shown in Figure 2.9. Agriculture in these areas is sensitive to non-climate drivers as well, such as market forces and political pressures (Montana et al., 2016 <sup>[[#fn:r354|354]]</sup> ; Sietz and Feola, 2016 <sup>[[#fn:r355|355]]</sup> ; Figueroa-Armijos and Valdivia, 2017 <sup>[[#fn:r356|356]]</sup> ) and shifts in water governance (Rasmussen, 2016). The majority of the adaptation activities are autonomous, though some are planned or carried out with support from national governments, non-governmental organisations (NGOs), or international aid organisations. Though many studies report on benefits from these activities which accrue to community members as increased harvests and income, systematic evaluations of these adaptation strategies are generally lacking. A range of factors, discussed below, place barriers which limit the scale and scope of these activities in the mountain agricultural sector, including a lack of finance and technical knowledge, low adaptive capacity within communities, ill-equipped state organisations, ambiguous property rights and inadequate institutional and market support ( ''medium evidence, high agreement'' ). Section 2.3.7 examines two other responses to decreasing irrigation water: wage labour migration, which often serves as an adaptation strategy, and displacement of entire communities, an indication of the limits to adaptation — this displacement is also due in some cases to natural hazards.

To cope with the reduced water supplies, planted areas have been reduced in a number of different places in Nepal (Gentle and Maraseni, 2012 <sup>[[#fn:r358|358]]</sup> ; Sujakhu et al., 2016 <sup>[[#fn:r359|359]]</sup> ). Adaptation responses within irrigation systems include the adoption of new irrigation technologies or upgrading existing technologies, adopting water conservation measures, water rationing, constructing water storage infrastructure, and change in cropping patterns (Rasul et al., 2019 <sup>[[#fn:r360|360]]</sup> ; Figure 2.9). Water delivery technologies which reduce loss are adopted in Chile (Young et al., 2010 <sup>[[#fn:r361|361]]</sup> ) and Peru (Orlove et al., 2019 <sup>[[#fn:r362|362]]</sup> ). Similarly, greenhouses have been adopted in Nepal (Konchar et al., 2015 <sup>[[#fn:r363|363]]</sup> ) to reduce evapotranspiration and frost damage, though limited access to finance is a barrier to these activities. Box 2.3 describes innovative irrigation practices in India. Local pastoral communities have responded to these challenges with techniques broadly similar to those in agricultural settings by expanding irrigation facilities, for example, in Switzerland (Fuhrer et al., 2014 <sup>[[#fn:r364|364]]</sup> ). In addition to adopting new technologies, some water users make investments to tap more distant sources of irrigation water. Cross-Chapter Box 3 in Chapter 1 discusses such efforts in Northern Pakistan, where landslides, associated with cryosphere change, have also damaged irrigation systems.

The adoption of new crops and varieties is an adaptation response found in several regions. Farmers in northwest India have increased production of lentils and vegetables, which provide important nutrients to the local diet, with support from government watershed improvement programs which help address decreased availability of irrigation water, though stringent requirements for participation in the programs have limited access by poor households to this assistance (Dame and Nüsser, 2011 <sup>[[#fn:r365|365]]</sup> ). Farmers who rely on irrigation in the Naryn River basin in Kyrgyzstan have shifted from the water intensive fruits and vegetables to fodder crops such as barley and alfalfa, which are more profitable. Upstream communities, with greater access to water and more active local institutions, are more willing to experiment with new crops than those further downstream (Hill et al., 2017 <sup>[[#fn:r366|366]]</sup> ). In other areas, crop choices also reflect responses to rising temperatures along with new market opportunities such as the demand for fresh vegetables by tourists in Nepal (Konchar et al., 2015 <sup>[[#fn:r367|367]]</sup> ; Dangi et al., 2018 <sup>[[#fn:r368|368]]</sup> ) and the demand for roses in urban areas in Peru (SENASA, 2017 <sup>[[#fn:r369|369]]</sup> ). Indigenous knowledge and local knowledge (Cross-Chapter Box 4 in Chapter 1), access to local and regional seed supply networks, proximity to agricultural extension and support services also facilitate the adoption of new crops (Skarbø and VanderMolen, 2014 <sup>[[#fn:r370|370]]</sup> ).

Local institutions and embedded social relations play a vital role in enabling mountain communities to respond to the impacts of climate driven cryosphere change. Indigenous pastoral communities who have tapped into new water sources to irrigate new areas in Peru have also strengthened the control of access to existing irrigated pastures (Postigo, 2014 <sup>[[#fn:r371|371]]</sup> ) and Bolivia (Yager, 2015 <sup>[[#fn:r372|372]]</sup> ). In an example of indigenous populations in the USA, two tribes who share a large reservation in the Northern Rockies rely on rivers which receive glacier melt water to irrigate pasture, and maintain fisheries, domestic water supplies, and traditional ceremonial practices. Tribal water managers have sought to install infrastructure to promote more efficient water use and protect fisheries, but these efforts have been impeded by land and water governance institutions in the region and by a history of social marginalisation (McNeeley, 2017 <sup>[[#fn:r373|373]]</sup> ).

High mountain communities have sought new financial resources from wage labour (Section 2.3.7), tourism (Mukherji et al., 2019 <sup>[[#fn:r374|374]]</sup> ) and government sources to support adaptation activities. Local water user associations in Kyrgyzstan and Tajikistan have adopted less water intensive crops and reorganised the use and maintenance of irrigation systems, investing government relief payments after floods (Stucker et al., 2012 <sup>[[#fn:r375|375]]</sup> ). Similar measures are reported from India and Pakistan (Dame and Mankelow, 2010 <sup>[[#fn:r376|376]]</sup> ; Clouse, 2016 <sup>[[#fn:r377|377]]</sup> ; Nüsser and Schmidt, 2017 <sup>[[#fn:r378|378]]</sup> ), Nepal (McDowell et al., 2013 <sup>[[#fn:r379|379]]</sup> ) and Peru (Postigo, 2014 <sup>[[#fn:r380|380]]</sup> ). In contrast, fewer adaptation measures have been adopted in Uzbekistan, due to low levels of capital availability and to agricultural policies, including centralised water management, crop production quotas and weak agricultural extension, which limit the response capacity of farmers (Aleksandrova et al., 2014 <sup>[[#fn:r381|381]]</sup> ).

Lowland agricultural areas which receive irrigation water from rivers fed by glacier melt and snowmelt are projected to face negative impacts in some regions ( ''limited evidence, high agreement'' ). In the Rhone basin in Switzerland, many irrigated pasture areas are projected to face water deficits by 2050, under the A1B scenario (Fuhrer et al., 2014 <sup>[[#fn:r382|382]]</sup> ; Cross Chapter Box 1 in Chapter 1). For California and the southwestern USA, a shift to peak snowmelt earlier in the year would create more frequent floods, and a reduced ability of existing reservoirs to store water by 2050 under RCP8.5 (Pagán et al., 2016 <sup>[[#fn:r383|383]]</sup> ) and by 2100 under RCP2.6, RCP4.5 and RCP8.5 (Pathak et al., 2018 <sup>[[#fn:r374|374]]</sup> ). The economic values of these losses have been estimated at 10.8 – 48.6 billion USD by around 2050 (Sturm et al., 2017 <sup>[[#fn:r385|385]]</sup> ). A similar transition to runoff peaks earlier in the year by 2100 under RCP2.6, RCP4.5 and RCP8.5, creating challenges for management of irrigation water, has been reported for the countries in central Asia which are dependent on snow cover and glaciers of the Tien Shan (Xenarios et al., 2018 <sup>[[#fn:r386|386]]</sup> ). In India and Pakistan, where over 100 million farmers receive irrigation from the Indus and Ganges Rivers, which also have significant inputs from glaciers and snowmelt, also face risks of decreasing water supplies from cryosphere change by 2100 (Biemans et al., 2019 <sup>[[#fn:r387|387]]</sup> ; Rasul and Molden, 2019 <sup>[[#fn:r388|388]]</sup> ).

<div id="section-2-3-1-3key-impacts-and-vulnerability-block-3" class="box"></div>

<span id="box-2.3-local-responses-to-water-shortage-in-northwest-india"></span>