Adaptive capacity of ski resorts in Western Norway to projected changes in snow conditions

ABSTRACT

Skiing is an important part of Norwegian culture, but as climate change leads to warmer, more variable winters, the ski industry needs to adapt. Despite the growing literature on climate change impacts on ski tourism, adaptation options, particularly beyond snowmaking, barriers and the financial and visitor experience implications remain under-researched. Employing projections for future snow and snowmaking conditions, this paper investigates adaptive capacity among seven Western Norwegian ski resorts. Adaptive capacity is examined in terms of physical situation, technology, economic resources, innovative ability, networks and institutions. We find that smaller resorts make up for poor economic performance by drawing on local community support and by implementing innovative efforts to diversify income. Nonetheless, despite high adaptive capacity with respect to networks, institutions and innovative ability, increased snow production costs will make operations in three low-lying resorts unviable as early as the 2030s, with salient implications for winter tourism patterns, small community economies and future participation in the sport. The results also suggest that studies using snow production model projections that represents physical and technical adaptive capacity only, may be conservative in their estimated impact of future climate change.

KEYWORDS:

• Climate change
• adaptation
• winter tourism
• Norway
• adaptive capacity
• ski industry

Introduction

Skiing is considered the most climate change-exposed tourism market, with ski operations in many regions already impacted by challenging climatic conditions (Steiger et al., 2019). Contemporary climate changes and lack of snow are making it more difficult for ski resorts to attract customers, reducing ski seasons and the number of operational ski areas in some markets (Scott et al., 2020a, 2020b; Steiger et al., 2019). The world’s mountain regions are losing on average five days with snow cover per decade, and the snow cover is projected to be reduced by 25% from the recent past (1986–2005) to the near future (2031–2050) (Intergovernmental Panel on Climate Change [IPCC], 2019).

Snowmaking is a key climate adaptation strategy to improve marginal snow conditions and extend the ski season, but it requires sufficiently low temperatures, which are declining in most mountain regions (IPCC, 2019). Even with advanced snowmaking capacity, large impacts on ski seasons are projected in several major ski tourism markets, particularly under higher-emission scenarios (Fang et al., 2019; Scott et al., 2020a, 2020b; Steiger & Abegg, 2018; Steiger and Scott, 2020). Many ski resorts in these diverse regional markets would have to double or triple snow production.

Snow reliability is one of the most important factors for destination choice (Steiger et al., 2020). Marginal snow conditions potentially lead to substantial demand shifts (Rutty et al., 2015) within a region or to other markets (Falk & Hagsten, 2018). Demand losses for individual ski areas have been found highest when substitute ski areas with better snow conditions exist (Steiger et al., 2020). Most climate change vulnerability assessments of ski tourism take a top-down approach, downscaling climate scenarios to estimate impact on the ski season length and the potential for technical snow production but overlooking adaptation strategies for ski resorts and skiers (Abegg et al., 2007; Rutty et al., 2015; Scott & McBoyle, 2007; Steiger et al., 2019). Increasingly, studies examine how climate change might impact demand and how supply and demand might respond to changing snow conditions (Demiroglu et al., 2018; Pons et al., 2014). Among tourism stakeholders, tourists have the highest adaptive capacity by changing the destination, timing or activity, if conditions are unacceptable for the desired activity at the desired destination (Scott et al., 2012). In their review of skier adaptation studies, Steiger et al. (2019) found that the most common adaptive response by ski tourists to perceived lack of snow would be spatial substitution (i.e. skiing at a different location than originally planned). Dawson et al. (2009) and Rutty et al. (2017) found that the observed decline in demand (skier visits) in snow-deficient and warm winters in Eastern North America was far less than survey-based studies had predicted. Generally, demand decline during anomalously warm winters ranged between 10% and 15% in North American and European markets; but was as high as 38% in Australia (Steiger et al., 2019).

Despite skiing’s cultural prominence in Norway, few studies have investigated potential climate change effects on the ski industry, future participation in the sport and the implications for winter tourism. Scott et al. (2020a) project that future natural ski season lengths in Norway could be shortened by 14–50% as early as the 2030s (depending on the region and the emission pathway) but that advanced snowmaking capacity would limit season losses in the same period to 4–15%. Even with advanced snowmaking, substantial shortening of the ski season is projected for regions of Norway under high-emission futures (−12% to −27% in the 2050s and −18% to 60% in the 2080s). The only demand-side analysis in Norway found that summer skiers were inclined to substitute skiing with other activities if snow conditions deteriorated (Demiroglu et al., 2018). Norwegian skiers’ potential response to marginal winter conditions remains uncertain, and how ski industry stakeholders perceive climate risk is unknown. Research from the Alps suggests that climate risks for skiing are not an important concern for stakeholders (Trawöger, 2014), but this may have changed with the increased prominence of climate policy and social momentum for climate action in European countries over the latter half of the 2010s.

Because of the limited research on climate change and skiing in Norway, particularly with respect to the adaptive capacity and potential response of skiers and ski operators, this study with industry stakeholders was exploratory in nature. This study is the first to examine the multi-faceted dimensions of ski operators’ adaptive capacity through case studies of seven ski resorts in Norway. In addition to assessing the ability to expand snow production, the study examines how determinants, such as networks, innovative ability, institutions and economic resources, can enable investments in snowmaking equipment and lifts as well as diversify income. The capacity of these adaptations to cope with projected climate change and maintain the viability of the seven ski areas are then discussed.

Assessing adaptive capacity in ski resorts

Adaptive capacity is commonly defined as the ability to cope with or adjust to changing climate conditions (e.g. Smit & Wandel, 2006). Because adaptive capacity is not directly observable, it is usually assessed and analyzed through proxy indicators (Engle, 2011). Some of these, such as economic resources, can be translated into quantifiable indicators, while others might be harder to quantify, for instance social capital (e.g. the dedication of a community to keep a local ski area operating, even at a financial loss). A much-cited chapter in the third assessment report of the IPCC defines adaptive capacity in terms of different determinants (economic resources, technology, equity, information, institutions and infrastructure) (Smit & Pilifosova, 2001). How the unit or system of interest has coped with previous stresses is also a way to assess adaptive capacity (Engle, 2011). The framework for assessing adaptive capacity in this study is inspired both by Smit and Pilisofova’s framework and newer works, by Keskitalo et al., (2011), Engle (2011), Juhola et al. (2012) and Dannevig et al. (2015). Table 1 summarizes the indicators utilized for each determinant of adaptive capacity.

Table 1. Framework for assessing ski resorts’ adaptive capacity.

Determinant of adaptive capacity	Indicators
Physical situation	Local climate, height above sea level and access to water for snowmaking.
Technology	Snowmaking equipment and ski lifts.
Economic resources	Turnover and profit for the last five years and owners’ financial resources.
Networks and collaboration	Cooperation with other industry actors at the destination, nationally and internationally.
Innovative ability	Developed new products, services and/or markets.
Institutions	Regulatory framework, including land use regulations.

Consistent with Smit and Pilifosova (2001), we include technology access and feasibility as a salient determinant of adaptive capacity. In the ski industry context, technology mainly relates to the ability to expand technical snow production and the potential to construct ski lifts in higher-elevation terrain. Infrastructure is a much-used determinant, but, in the case of ski resorts, it overlaps with technology in terms of what is salient for adaptation (i.e. ski lifts and snowmaking systems).

Economic resources, defined as ‘economic assets, capital resources, financial means, wealth, or poverty’ (Smit & Pilifosova, 2001), are a central determinant of adaptive capacity in many studies (Dannevig et al., 2015; Juhola et al., 2012; Keskitalo et al., 2011). The ability to diversify economically and exploit different income sources has been found to increase adaptive capacity in many studies in various sectors and at different scales (e.g. Dannevig et al., 2015; Juhola et al., 2012; Keskitalo et al., 2011). In this study, economic resources translate directly to a ski resort’s ability to invest in increased snowmaking capacity, extend the ski area into higher terrain or develop new products to contribute to revenue diversification (e.g. Scott & McBoyle, 2007).

The physical situation and climatic conditions can be assets that provide adaptive capacity, and resemble natural capital, as included in other adaptive capacity studies (e.g. Dannevig et al., 2015; Kofinas et al., 2013).

Networks and collaboration represent social capital, which are resources and services an actor can draw on (Coleman, 1988), and the ability for collective action (Adger, 2009). Social capital or networks, which are included in several tourism resilience frameworks (Bec et al., 2016; Calgaro et al., 2014; Wyss et al., 2017) and the adaptive capacity framework by Eakin and Lemos (2006), are also relevant, as ski resorts tend to belong to a broader tourism industry in their respective destinations.

The adaptive capacity literature tends to downplay the importance of human agency (Dannevig et al., 2015; Posch et al., 2020; Wyss et al., 2017), but, as Westley et al. (2013) note, coping with system change requires strategic agency, entrepreneurship and innovation. These properties rest with individuals, and motivated and engaged individuals’ agency has been found to be crucial in mobilizing adaptation both in governance institutions (Dannevig et al., 2013) and nature-based tourism operators (Dannevig et al., 2015; Tervo, 2008). While some frameworks include these properties under institutions, as they shape and guide human agency (Keskitalo et al., 2011; Smit & Pilifosova, 2001), in this analysis, the importance of agency and entrepreneurship is emphasized with a separate determinant called ability for innovation. We assess this in terms of past and current actions to diversify income sources or mitigate losses from poor snow conditions by developing new products, services or markets.

Institutions, as defined in Smit and Pilifosova (2001), are a broad determinant encompassing management, strategies and legislation. Later frameworks split this determinant into social capital (Eakin & Lemos, 2006) and the ability to cope with previous hazards (Engle, 2011; Keskitalo et al., 2011). We consider institutions in terms of how regulatory frameworks enable or hamper adaptative actions and innovations. For ski resorts, land use management, environmental assessments and zoning plans determine the potential for upward expansion of ski lifts, utilization of rivers and lakes (versus dedicated reservoirs) as water supply for snow production and regulations on slope design (i.e. smoothing), the construction of pipes and other infrastructure.

Methods

Consistent with common approaches to adaptation assessments, this analysis combined top-down and bottom-up approaches (e.g. Smit et al., 2010). The top-down component utilized the SkiSim model (Scott et al., 2020a) to develop projections for snow conditions and ski operations at the case study ski resorts. The bottom-up component used a multi-method approach, including a survey to inform ski industry stakeholder about the ski season and snowmaking projections from the SkiSim modelling, interviews with ski industry stakeholders about their adaptation strategies and capacities, and a joint workshop with ski industry stakeholders and the researchers on adaptation options and barriers.

Case study ski areas

The study included seven ski areas in Sogn og Fjordane, Hordaland and Buskerud counties in Norway (see Table 2). These counties have 12, 13 and 30 ski areas, respectively. For this study, we selected the ski areas in and around former Sogn og Fjordane county (which during the time of writing was merged with Hordaland county, and is now called ‘Vestland county’), which constitute the core of Norway’s Fjord region (see Figure 1). We selected one resort outside the Fjord region, as it is 40 min’ drive from the county border that demarcates the Fjord region and therefore attracts many skier visits from this region.

Figure 1. Ski Areas Participating in the Study.

Winter tourism has been increasing in the Fjord region of Norway in the last decade. Regional destination organizations are working on making the fjord areas year-round destinations, but winter visitor numbers are still a fraction of the summer season’s numbers in terms of overnight stays. In 2016, the municipality of Voss, where Myrkdalen ski resort is located, had 1,07,000 overnight stays from October to April; foreigners constituted just 14,000 (13%) of these. The summer season (May to September), by comparison, had 1,73,000 stays, of which foreigners constituted 99,000 (57%).

The largest ski resorts in this regional market are Hemsedal and Myrkdalen (see Table 2). Myrkdalen, the biggest ski resort in Western Norway, has grown rapidly since opening in 2002, both in turnover and visitor numbers. A new chairlift in 2013/2014 doubled the capacity of the ski centre. Hemsedal is located right between Norway’s two largest cities, and it attracts visitors from both Western and Eastern Norway. For years it has been Norway’s second biggest ski resort in terms of turnover, number of visitors and number of lifts. The other ski resorts in this study are small, primarily catering to regional and local markets, although Sogndal has a reputation as a ‘powder destination’ and receives visitors from all over Scandinavia (between 700 and 1000 winter time overnight stays from Scandinavian countries at the resort).

Table 2. Characteristics of ski resorts involved in study.

Ski area	County	Lifts (n)	Bottom station height above sea level	Top station height above sea level
Harpefossen	Sogn og Fjordane	6	242	820
Stryn	Sogn og Fjordane	4	325	995
Jølster	Sogn og Fjordane	4	250	780
Sogn	Sogn og Fjordane	4	325	994
Sogndal	Sogn og Fjordane	5	500	950
Myrkdalen	Hordaland	8	475	1060
Hemsedal	Buskerud	20	640	1450

Projections for future ski conditions at study area ski resorts

This study of adaptive capacity was part of a larger national study that generated projections for future snow conditions and ski seasons, including conditions with current and advanced snowmaking technologies, for 105 Norwegian ski resorts. Aggregated results for ski resorts in the five geographical regions of Norway were presented in Scott et al. (2020a). This analysis provided detailed modelling at the seven case study locations to inform the analysis of differential adaptive capacity. The SkiSim model employs time series data for snow depths, temperature and precipitation over a 30-year baseline period (1981–2010) from weather stations near the ski resorts. Climate data were provided by the Norwegian Meteorological Institute. Meteorological stations were chosen based on distance to the ski area, similar elevation and data completeness. Some adjustments to the choice of the most representative weather stations were made based on the ski resorts’ knowledge of local weather conditions.

Climate change scenarios (monthly) for the study areas were derived from regionally downscaled temperature and precipitation projections from the EURO-CORDEX regional climate model and obtained from the Norwegian Climate Service Centre (see Hanssen-Bauer et al., 2015). These greenhouse gas emission scenarios represent two representative concentration pathways (RCPs) defined by the IPCC: a moderately optimistic scenario largely consistent with successfully achieving country emission reduction pledges to the Paris Climate Agreement (RCP 4.5) and a high-emission scenario (RCP 8.5). The monthly projections were downscaled to daily temperature and precipitation inputs into the SkiSim model using the LARS weather generator (Semenov & Stratonovitch, 2010; see Steiger, 2010 for additional details of the SkiSim model). Ski operations simulations were made for three periods: 2030s (2021–2050), 2050s (2041–2070) and 2080s (2071–2100). Projections were developed for the base station elevation (the lifts’ starting point) and for any ‘middle stations’, defined as parts of the ski area that are reachable by car or chairlift that could serve as temporary base stations when poor snow conditions exist at lower elevations.

As a benchmark, the ‘100-day rule’ was used to distinguish between snow-reliable and non-snow-reliable ski resorts. If the skiing season lasts at least 100 days in 70% of winter seasons, the ski area is considered snow-reliable (as per Abegg, 1996). This performance indicator has been used in several studies on climate change impacts on ski resorts (Scott et al., 2015; Scott et al., 2020a, 2020b; Steiger & Abegg, 2018), and our stakeholders did not suggest any other heuristic of financial viability. Having skiable conditions during the economically important Christmas holiday season has been used in the aforementioned studies as an additional economic indicator and was confirmed as extremely important by the case study resorts. The SkiSim model was used to project the likelihood of the case study ski areas being operational in late December and early January.

The projections of season length, Christmas holiday operations and snowmaking requirements were presented to the ski resort representatives. In most cases, the weather stations used for meteorological data to represent the study areas’ climate conditions were discussed. Particularly for Myrkdalen, there was concern that the projections did not reflect the local climatic conditions, which are well known in the region to be colder and more snow-abundant than places at the same altitude. The potential for micro-climate conditions at the ski areas is not captured in the projections due to insufficient weather data. Stakeholders from the other resorts found the baseline simulations mostly agreed with their experiences, which provided important local validation of model outputs, and provided them a valuable tool to explore changing climate risk in the future.

Data collection with industry stakeholders

Semi-structured interviews (n = 17) were conducted in all case study destinations to determine current responses to poor snow conditions, explore industry views on the potential for an increase in tourists choosing Norwegian ski centres over less climate-reliable destinations, and analyse adaptive capacity. Individual semi-structured interviews (n = 3) were conducted with the Hemsedal operations director, the managing director of Jølster and the chairman of the board for Jølster. The other 14 interviews were conducted as three group interviews due to the ski resorts’ close proximity and membership of the same destination marketing organizations, which also took part. Group interviews are beneficial to get conversation going between interviewees, whereby ‘people respond to each other’s views and build up a view from the exchange that takes place in the group’ (Bryman, 2012, p. 501). The downsides are that some participants might not contribute much to the discussion, and transcription is more challenging (Bryman, 2012). The same interview guide was employed for all interviews, with the single interviews typically lasting 90 min and the group interviews taking two hours each. The group interviews were organized as follows: (1) Sogndal and Sogn resorts were done as one interview with the managers and a board member from each resort and one representative from the regional destination marketing organization (DMO) (n = 5); (2) Myrkdalen included three directors (hotel, marketing and finance) and the managing director (n = 4); and (3) Stryn included the manager, three former and current board members and one regional DMO representative (n = 5). SkiSim projections were presented to the stakeholders, and they were asked to comment on the projections and provide feedback based on their experiences. The interviews also discussed the consequences of and adaptation to recent seasons with little or poor snow, including conversations about existing noticeable changes in climatic conditions and the effects on the ski centre’s operations and any further development. The interviews also discussed the destination’s adaptive capacity in terms of the following: physical situation and experience of local climatic conditions and change; access to snowmaking technology and ability to expand snowmaking capacity and expand ski terrain upwards; networks and collaboration; economic resources; and innovative ability.

Document analysis was also used to review the ski resorts’ financial records and the land use regulations. Existing data on season length, ski slope length, lift capacity and base and summit height were incorporated into the SkiSim model, allowing for adjustment from the meteorological station’s altitude to the ski resort’s base station and summit (using local lapse rates). Data were also collected on profits and ownership. We reviewed financial records for the last five years and ownership structure from each company’s reports,¹ which we applied as indicators of economic resources. The number of overnight stays in the winter season in the resorts have been fetched from the Norwegian tourism and travel database ‘Statistikknett²’. We reviewed land use plans³ with regulations to assess potential for expansion under the ‘institutions’ determinant.

Survey. The results from a survey sent to all Norwegian ski resorts (n = 105; 85% response rate) we used for the seven ski areas involved in this study. We extracted information on season length (number of days), snowmaking capacity (percentage of slope covered), snowmaking chemicals and costs and decision-making on when to start and stop snowmaking. All six resorts with snowmaking capacity in the study area responded to the national survey that was used to inform national modelling reported on by Scott et al. (2020a).

Climate change impacts in the case study region

The study area in the central part of Western Norway is projected to experience a temperature increase of 2.3°C by the 2050s under a high-emission scenario (RCP 8.5) (relative to the 1971–2000 reference period) (Hanssen-Bauer et al., 2015). Precipitation increases in all seasons by mid-century, and for Western Norway it has already increased by 16% since 1900. For central Western Norway, winter precipitation is projected to continue to increase by 4% in the 2050s (relative to the 1971–2000 period) but with substantial uncertainty (Hanssen-Bauer et al., 2015).

The modelling of future conditions for skiing show two marked trends: (1) Western Norway will be significantly more exposed to climate change-induced ski season losses; (2) inland areas will still enjoy relatively stable winters even by the end of the century (Scott et al., 2020a). Snowmaking can offset deeper losses in skiable days. To preserve ski operations of at least 100 days per season, technical snow production increases by 31% under the RCP 4.5 scenario in the 2030s, 55% in the 2050s and 90% in the 2080s (Scott et al., 2020a). For the higher-emission, warmer RCP 8.5 scenario, snowmaking requirements increase 41% in the 2030s, 108% in the 2050s and 259% in the 2080s.

Adaptive capacity of the ski resorts

Although the ability to install or expand snowmaking is the most common adaptive measure for countering a less snow-secure future, as Scott and McBoyle (2007) and others have pointed out, it is not the only adaptive strategy. The determinants of adaptive capacity used in this study (see Table 3) are each presented below. The study area scale SkiSim modelling results on snowmaking needs or operating conditions are incorporated in the discussion where appropriate.

Table 3. Indicators for adaptive capacity determinants in the case study ski resorts.

Resort	Physical situation	Technology (Likely to achieve needed snowmaking by 2050)	Economic resources R = Average annual revenue in 1000 NOK; Average growth rate for the last five years in % P= Annual average profit in 1000 NOK; average % of revenue for the last five years	Innovative ability	Networks and collaboration	Institutions Land use regulations allow for increased snow production and upward lift expansion
Harpefossen	Low-lying. Road access to middle station.	No.	R = 4402; 11% P = −259; −9%	Introduces cat skiing.	Part of regional ski pass and local support.	Yes
Stryn	Low-lying.	Maybe.	R = 2991; 1% P = 86; 1%	Destination development ongoing.	Part of regional ski pass and local support.
Jølster	Low-lying. Road access to middle station.	No	R = 2876; 3% P = 86; −4%	–	Local support	Yes
Sogn	Low-lying.	Maybe.	R = 3928; 1% P = −42; −2%	–	Part of regional ski pass and local support.	Yes, but low potential for expansion.
Sogndal	Large annual snowfall.	Yes, but requires new reservoir.	R = 7166; 25% P = 120; −4%	Comprehensive destination development. Brand development. Ski touring.	Part of regional ski pass. Local guide companies and local support.	Yes.
Myrkdalen	Large annual snowfall. Chairlift to middle station.	Yes, but requires additional water supply.	R = 141415; 19% P = 1587; 11%	Comprehensive destination development and attracts foreign customers.	Engine in local tourism development. Local guide companies. International tourism operators.	Yes, but conflicts.
Hemsedal	Chairlift and road to middle station.	Yes.	R = 502423; 7% P = 13913; 28%	Took control of lodging and hotel booking and attracts foreign customers. Snow guarantee.	Engine in local and regional tourism development. International tourism operators.	Very limited potential for expansion. Water supply is sufficient within existing zoning plan.

Table 4. Snow production coverage and capacity in the resorts.

Resort	Slope covered (m)	% coverage of technical snow production	Snow production capacity (number of days with continuous snow production to open the ski area without natural snow)
Harpefossen	1500	15	6
Stryn	600	15	14
Jølster	5300	90	3
Sogn	1000	60	6
Sogndal	0	–	–
Myrkdalen	5000	18	5
Hemsedal	19100	42	7

Physical situation

Although all but one ski area in this project is in Western Norway, their physical situations differ greatly. This determines the extent to which a ski resort needs to adapt to deteriorating snow conditions, how feasible snowmaking capacity expansion is and the potential for expanding ski areas into higher, more snow-secure terrain. Base station altitude ranges from 220 metres above sea level (masl) at Jølster to 640 masl in Hemsedal. These two centres also have the lowest and the highest top stations (780 and 1,450 masl, respectively). The ski areas also differ in how strongly they are affected by the coastal climate, with Hemsedal located further inland than the other centres.

Local climatic conditions lead to highly diverse impacts for the ski centres. Myrkdalen, Sogn and Sogndal all receive more snow than surrounding areas. Looking at the probability of skiing on at least 100 days with snowmaking (see Figure 2), Hemsedal is the only centre that fulfils the 100-day rule with snowmaking in 2080 under a high-emission scenario (RCP 8.5). Sogn and Sogndal will get sufficient natural snow most years well into the middle of the century, and, along with Myrkdalen, will also manage 100 days with snowmaking under the moderate-emission scenario (RCP 4.5). Neither Harpefossen nor Jølster and Stryn with their current lifts are projected to be viable under any emission scenario after the 2030s, despite snowmaking.

Figure 2. Probability of 100-day seasons with snowmaking.

All the resorts have lifts starting at higher elevations than the base station, but only Myrkdalen, Hemsedal, Harpefossen and Jølster have road or chairlift access to these lifts. Jølster and Harpefossen could still offer skiing in their resorts’ higher parts beyond 2030, when the lower part becomes unviable. The resort’s physical situation and economic resources and land use regulation (institutions) determine whether expanding the resort upwards is an option. Sogndal, Jølster, Harpefossen and Stryn all have physical situations that allow upward expansion. When confronted with snow conditions projections (Figure 2), the Jølster ski resort manager expressed a desire to move the entire resort upwards in the long term.

Technology

All the ski resorts wanted to expand technical snowmaking, which is the most important adaptation measure in the industry world-wide (e.g. Scott et al., 2012; Steiger et al., 2019). For the resorts with snowmaking technology, the coverage of skiable terrain ranges from 15% in Stryn and Harpefossen to 90% in Jølster (see Table 4). However, as illustrated in Figures 2 and 3, this will not help Jølster overcome the impact of climate change. Figure 3 shows the required amount of produced snow the resorts need to produce to maintain skiable conditions through the winter season. As Figure 3 shows, will the amount of produced snow decline for Jølster, Stryn and Harpefossen. SkiSIM model projections assume 100% snow production coverage, meaning that in addition to increasing production with current installed capacity, the resorts would also need to expand coverage to produce the required amount as illustrated.

Figure 3. Modelled snowmaking requirements in cm.

The resorts with no or low snowmaking capacity are particularly vulnerable to a late winter onset. Sogndal, the only resort lacking snowmaking equipment, occasionally loses the important Christmas week due to lack of snow. However, according to the projections, they will have sufficient natural snow conditions until the 2030s and thus a decade to raise the capital to install snowmaking equipment. Hemsedal has medium snowmaking coverage but can already offer sufficient snow to start the season in November most years. Myrkdalen, the other large resort in the study, had lower coverage and expressed the need to expand coverage to ensure the ability to operate with poor natural snow conditions, particularly to start the season in early December. Harpefossen and Jølster have experimented with snowmaking that includes the chemical ‘Snowmax’ to produce snow at warmer temperatures. However, the technology did not prove cost-effective for the production of snow in conditions above 0°C due to high financial costs.

Another technical adaptation is to create new skiing terrain at higher altitudes. Both Harpefossen and Jølster, which have struggled with variable winters for many years, have expanded their facilities with new lifts starting at higher altitudes at around 500 masl in 2010. Jølster has recently built an additional lift at the middle station to expand the skiable area upwards to more snow-secure terrain and wants to improve access to their middle station with better roads and parking facilities The projections for Harpefossen show 126 days at the base station (250 masl) and 150 days at 500 masl, highlighting the importance of shifting to higher elevations where feasible. This gap increases substantially when looking at future scenarios. The high-emission scenario has a 67-day difference between the base (250 masl) and the middle station (500 masl) already in the 2030s. In the 2080s, projections for Harpefossen show a decline to four days at base station when using the high-emission scenario, while middle station projections show a season of 36 days.

The stakeholders from all the resorts, except Myrkdalen and Sogndal, stated that the projected increases in snow production by the end of the century would be unrealistic to achieve with current snow production technology, due to the high cost of producing snow and expanding snow production coverage. As the Harpefossen representative said: ‘If the projections for the 2050s are correct, we will just have to close’. An increase in snow production would require major investment in infrastructure (e.g. snow guns, water pumps, etc.) and require access to water. Only Hemsedal has access to sufficient water for expanding their coverage and production significantly, as several smaller lakes are situated above the main resort. The other resorts would have to build reservoirs at elevation and install additional pumps and pipes. When snow production operates, it requires much electricity, and for Jølster it was already challenging to pay the cost of electricity after longer periods of snow production, according to its manager. As Figure 3 shows, snowmaking will be unfeasible due to warm winter temperatures in the three low-lying resorts of Jølster, Stryn and Harpefossen by the 2030s. Even if they operated from their middle stations, the stakeholders deemed the cost of increased snowmaking (in terms of increased slope coverage and increased volume of snow produced) unfeasible.

Economic resources

All the resorts viewed economic resources as a key barrier to expanding snowmaking capacity and increasing snowmaking with current infrastructure capacity. The resorts’ economic resources vary widely, which is partly a function of their size and the market situation. We reviewed the economic performance (revenue and profit) from 2015 to 2019 (see Table 3) and found that the five smaller resorts experienced large fluctuations in revenue and were not profitable in two to three years out of the five-year period. The largest resort, Hemsedal, is the most profitable, with an average annual profit of 28%. Myrkdalen has also made solid annual profit (11% on average), despite substantial concurrent capital investments in infrastructure. According to the stakeholders, snow conditions are the most important revenue-determining factor, and years with poor economic performance are often due to poor snow conditions and shorter seasons. Sogndal is the only smaller resort that has been able to increase revenue and profit in the last three years. The smaller resorts are able to continue operation and even continue to invest in lifts and snow production equipment because of large voluntary work efforts by local sports clubs, sponsorship by local businesses, and direct financial support from the local municipalities, which together with the local sports clubs own most of the shares in the five smaller resorts.

Hemsedal is owned by a Swedish publicly listed company, which owns several ski resorts in Scandinavia and one in Austria. This company is able to fund large investments, but decisions are taken in the company headquarters in Sälen, Sweden, not in Hemsedal. If the company wanted to expand snowmaking capacity, it can, according to our informant, but only if it is certain that it will get a return on the investment. Myrkdalen is also owned by industry investors. The other ski resorts are to a large extent owned by municipalities and sports clubs, but Sogndal has also seen the arrival of professional industry investors, who have contributed minor investments in new lifts and produced plans for a huge expansion of the resort with hotels and gondolas. According to our industry informants, these investors did not consider climate change risks when making decisions in either resort.

Innovative ability

The resorts have, to various degrees, displayed an ability to develop products, services and markets to increase their profitability and long-term viability, which signifies an ability for innovation and entrepreneurship (e.g. Dannevig et al., 2015). If successful, these innovations could also improve the resorts’ economic viability, offsetting losses from poor snow conditions. A common strategy for ski resorts to increase lift ticket sales/skier visits (and increase the utilization rate) is to invest in accommodations. Hemsedal has taken steps to develop the local tourism lodging market, according to the its owner’s policy, including the acquisition of two hotels and the booking service for apartments and cabins. The same strategy has been applied by Myrkdalen. Both resorts receive their main revenue (see Table 2) from ski pass sales and not from associated services like food and lodging. While this strategy can increase ski pass sales, it is, according to our informants, highly uncertain if it can offset reoccurring poor snow conditions.

Hemsedal offers a ‘snow guarantee’ to its customers, whereby the resort offers refunds lift tickets if the number of open slopes falls under a certain level. This is particularly relevant for Hemsedal, as 60% of the overnight guests are foreigners and book their stay long before they know the snow condition at the resort. Sogndal has sought to market itself as a ‘powder and off piste’ destination and has introduced a bus to collect off piste skiers. While this indicates innovative ability, it is very vulnerable in the context of projected climate change, as many more days will be above 0°C, meaning wet snow instead of powder snow.

The two large resorts, as well as Sogndal, are developing summer products, mainly mountain biking, to further diversify revenues. This is common strategy over the last 20 years for ski resorts to transition to four-season mountain resorts (Scott & McBoyle, 2007). However, according to the stakeholders, this was intended not to ensure greater ski lift profitability but to offer activities to cabin owners and tourists. Stakeholders were clear that summer products could not offset or compensate for losses in skier visits from a poor winter season.

Hemsedal and Myrkdal seek to attract foreign customers, particularly Swedish, Danish and British customers. In Myrkdalen, they have noted an increased interest from tour operators that usually operate in the Alps looking for more snow-secure destinations. Stakeholders agreed that deteriorating snow conditions in the Alps could benefit Norwegian resorts, similarly to what has occurred in Swedish Lapland (Demiroglu et al., 2019).

During the workshop, possible adaptive solutions to deteriorating snow conditions besides snowmaking were discussed (e.g. measures to increase resort attractiveness with different customer services, particularly family-friendly activities). The two large resorts offer ski activities for young children. However, to offset the lack of snow, stakeholders agreed that such measures were not enough. One suggestion was a swimming pool, as other resorts’ experiences suggested that these could compete with the ski slopes in popularity on poor weather days.

Networks and collaboration

All the resorts participated in various forms of collaboration and in networks with other tourism actors in their respective destinations. Such networks can contribute to increased adaptive capacity by channelling resources to the resorts (Wyss et al., 2017). As noted under economic resources, the five smaller resorts survived successive years of deficits and poor economic performance thanks to major voluntary efforts by their local community and economic support from sponsors and their owners. This clearly indicates social capital these resorts can draw on and which could also help them to some extent overcome more frequent years with poor snow conditions in the future.

Hemsedal’s relationship with its local community stands in contrast to the other resorts, as they are a significant tourism asset in their local community and municipality. A new and innovative collaboration was with a ferry company operating between Denmark and Norway. Together, they offered a ‘ferry and ski’ package to the Danish market. Both Myrkdalen and Hemsedal also collaborate with large international tour operators.

All the resorts, except Hemsedal, collaborated with property developers to facilitate cabin and accommodation construction. In Sogndal and Myrkdalen, a portion from real estate sales is set aside for construction of new ski lifts. This model has funded both resorts’ development and also exists in Jølster. Six of the seven resorts took part in agreements for regional ski passes with other resorts, which improves customer relationships for all of their visitors.

Institutions

Internal institutional factors include organizational structure and management. The five smaller resorts have only one to two full-time employees, including the manager, who reports to the company board. The organizational structure is thus quite simple, and the manager’s ability and competence are the major factor that influences internal institutional adaptive capacity. The situation is quite different for Hemsdal, which is part of a large company. Our informant in the company stated that all decisions about larger investments needed to be approved by the company management in Sälen, Sweden. The external institutional factor that to a large degree shapes adaptation options for ski resorts is land use regulations. Municipal land use plans govern land use in Norway, and if the ski resorts want to build higher-lying lifts, construct water reservoirs or pipe water from streams and lakes for snow production, it is necessary that the current land use regulations allow for this. As shown in Table 3, several resorts have secured approved zoning plans or formulations in the current land use plan for upward expansion. Sogndal, for instance, has a zoning plan for a new lift, while the municipal land use plan for Voss, where Myrkdalen is located, has allocated an area almost four times as large as the current resort for future lift and slope development. In Hemsedal, however, the municipality has adopted strict land use plan regulation in the area around the ski resort, so no further expansion is possible. This resort is already large (in a Norwegian context) with its 20 lifts, so may not plan to expand further. Nonetheless, these differences even within a small region of Norway, illustrate how adaptive capacity can be strongly influence at the local scale.

Summary of adaptive capacity determinants

Do the different non-technical determinants (see Table 3 for a summary) translate to adaptive capacity in the face of climate change? Physical situation determines the resort’s natural capital or ecosystem services it can draw on from the landscape, water access and local climate (e.g. Dannevig et al., 2015). It determines the extent to which technological adaptation measures are needed and feasible. This is the extent to which many studies in the literature are able to examine climate change risk (including the many regional studies using the SkiSim model). These technological solutions also hinge on economic resources, as it is costly to invest in and operate snowmaking equipment. While the resorts benefited economically from cooperation and networks by, for instance, attracting more foreign visitors and improving their economic results through innovations, such as diversification to other winter and summer activities, these emerged as less salient than technology, physical situation and economic resources, both according to the stakeholders and our assessment of the empirical material from the case studies. This contradicts results from Switzerland, where collaboration within a destination was found to offset negative impacts from climate change (Wyss et al., 2017). Similarly, in many North American destinations with ski villages largely owned by a single company or consortium of companies, collaboration is established at the project development stage and part of the business model. The notable exceptions are the smaller resorts that benefit from voluntary work and economic support from their local communities, which have proven crucial to survival in years with poor economic results. While innovations such as Hemsedal’s ‘snow guarantee’ in the short term might offset or postpone reduced demand from several successive winters with poor snow conditions, stakeholders from this resort were unsure how they could pay for the high cost of increased snowmaking projected for the 2030s and onwards. Current land use plans allow upward expansion in all resorts except Hemsedal, thus functioning as an institutional factor contributing to lower adaptive capacity at this ski area. Upward expansion could be crucial in Jølster and Harpefossen, as their lower parts are projected to become unviable for snow production after the 2030s.

Conclusion

There is no avoiding that snow availability is crucial for the viability of the resorts studied in this project. All types of collaboration and innovation cannot make up for the lack of this crucial physical asset. Stryn and the lower parts of Harpefossen and Jølster are likely to be unviable in the 2030s, unless new technology for less costly snowmaking above 0°C is introduced. Higher lifts and ‘middle stations’ that can be reached with either car or chairlifts could also offset deteriorating snow conditions for an uncertain period of time. Viable snowmaking and new infrastructure require sufficient economic resources, which is thus a main source of adaptive capacity. The abovementioned resorts have also struggled with poor economic performance (see Table 3), partly as a consequence of poor snow conditions, which further hamper the ability to invest in snowmaking equipment, despite support from their local communities. Thus, based on our analysis of adaptive capacity we find it unlikely that these three resorts will be viable after the 2030s. This is an important realization for local communities to plan for.

There is a risk that resorts are not assessing their regions’ climate change projections before investing in snowmaking capacity, which then could be rendered useless as temperatures exceed the limits for snow production. Huge snowmaking investment could therefore prove maladaptive as a form of ‘stranded asset’ in the long term for some of the resorts in this study area.

Stakeholders worried that the absence of natural snow would reduce their destinations’ attractiveness. Conversely, the more snow-secure resorts in the study might benefit from the deteriorating snow conditions in the Alps and receive ‘ski refugees’ from the rest of Europe (e.g. Demiroglu et al., 2019). This also illustrates the salience of the physical situation and local climatic conditions as sources of adaptive capacity and future ski tourism market dynamics. The salient determinants of adaptive capacity in this study are thus technology, physical situation, and economic resources.

A key finding from this study is that future research to better understand the implications of differential climate risk for destination competitiveness needs to develop methods to incorporate business model and economic capacity to assess which ski areas might be able to afford the substantial increases in snowmaking projected for most ski areas. The results suggest that studies using the SkiSim model that represents physical and technical adaptive capacity only, may be conservative in their estimated impact of future climate change. Assuming that all ski areas are able to afford state-of-the-art snowmaking technology and capacity, as done in many climate change impact assessments (Steiger et al., 2019), might lead to an overly optimistic view of the future, if some ski areas modelled as snow reliable are financially not capable of making necessary investments into snowmaking as revealed by our case study. If this regional study is at all representative of other ski tourism regions, similar analyses are needed to refine the SkiSim results based on the economic capacity of individual ski areas to support advanced snowmaking adaptation.

Acknowledgements

The research underlying this article is funded by the Regional Funding Programme for Tourism Research in Sogn and Fjordane County (2014–2019) and by the Research Council of Norway through the project ‘CLIM-TOUR’ (grant number 281006). The authors want to thank the industry stakeholders that took part in the project for their time and effort, as well as the editors and reviewers for constructive and helpful comments that has improved the quality of the article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The research underlying this article is funded by the Regional Funding Programme for Tourism Research in Sogn and Fjordane County (2014–2019) and by the Research Council of Norway through the project ‘CLIM–TOUR’ (grant number 281006).

Notes

1 http://www.proff.no

2 http://statistikknett.no

3 https://kommunekart.com