The most common snow indicator for ski tourism is the number of days with a certain amount of snow on the ground, the “snow days.” Different thresholds have been proposed, for example 1, 5, 15, 30, or 50 cm. The snow depth requirements are dependent on slope characteristics: the rougher the terrain, the more snow is needed. Most ski areas, however, operate on landscaped terrain, i.e., slopes have been technically altered to reduce the required snow depth and to facilitate grooming. Therefore, 30 cm is typically taken for ungroomed snow—a threshold widely accepted since its introduction by Eckel (1938 in Witmer 1986) in both the science and ski tourism communities. The lower thresholds (e.g., 5 cm) are used for the “white winter landscape” (Strasser et al. 2014) or the “wintry atmosphere” (Schmucki et al. 2017)—an indicator that is not directly related to the operation of a ski area but to the look and feel of the environment.
In some studies (e.g., Scott et al. 2006), a day with ≥ 30 cm of snow is called a “skiable day,” and the number of days with snow depths ≥ 30 cm the “ski season length” (also: number of days between the ski opening (Snow Water Equivalent (SWE) > 120 mm w.e. (kg m−2) after 1 November for at least 5 consecutive days) and the ski closing (SWE < 80 mm w.e. for at least 10 consecutive days between the opening and 30 April), Marke et al. 2015). Based on the operational practices of ski areas in Eastern Canada, Scott et al. (2003, 2006, 2007) suggested that ski areas were assumed to close if one of the following conditions occurred: snow depth less than 30 cm, maximum temperature higher than 15 °C, or 2-day liquid precipitation exceeding 20 mm. The same authors also point to the difficulty to determine the length of the ski season with meteorological criteria only. Ski area managers may start the operation early, i.e., before the desired snow depth is reached, because they have to compete with neighboring ski areas (or to cash in during high-demand periods), and they may end it early because the demand is lacking and costs can be reduced (Scott et al. 2007). This is often the case towards spring when people prefer summertime activities although there is still a lot of snow in the ski areas (Mayer et al. 2018).
Starting in the late 1970s, a special meaning was attributed to a particular number of snow days. Referring to French ski areas, Barbier (1978) wrote that 120 snow days are necessary to assure the profitability of ski area investments. Similarly, Witmer (1986: 193) stated that “in order to have an economically viable investment in [Swiss] winter sports areas, among others, the installed facilities need to be utilized for at least 100 days per season, which is only possible with a snow cover of sufficient thickness.” In Australia, “60–70 days is about the minimum for a viable downhill ski operation” (Galloway 1988: 428), and in North America, Mieczkowski (1990: 254) proposed that at least 100 (in the East) or 120 days (in the West) are necessary “to be commercially viable.” Witmer’s suggestion, commonly known as the 100-day rule, has become very prominent. Applications include, for example, Abegg (1996), König and Abegg (1997), Elsasser and Bürki (2002), Durand et al. (2009), Steiger and Stötter (2013), and Pons et al. (2015) for Europe, Bark et al. (2010) and Dawson and Scott (2010, 2013) for North America, and Hendrikx and Hreinsson (2012) for New Zealand. Scholars looking at Australian ski areas still refer to Galloway (e.g., Pickering 2011). However, the 100-day rule (or any other similar rule) cannot be considered a robust economic indicator (Abegg 1996, 2012). Economic success is dependent on sufficient snow at the right time (e.g., key periods), good weekend-weather, and many other non-climate-related factors (e.g., Elsasser and Bürki 2002). This has been known for long, but still, the 100-day rule is utilized as a proxy for the profitability, viability, and even sustainability of ski areas.
The “line of snow reliability,” combining the 100-day rule with altitude, goes back to earlier research in Switzerland. Föhn (1990) found that the criteria for the 100-day rule are met in areas above 1200 m. This line of snow reliability, he further suggested, will move up to 1500 m assuming a + 3 °C warming. Messerli (1990) used this information to draw a map, identifying a critical zone between 800 and 1500 m with ski areas being “at risk” or “highly at risk.” Abegg (1996) adapted Föhn’s approach and calculated the number of naturally snow-reliable ski areas under current and future climate conditions: a ski area was considered naturally snow reliable if the upper half of its altitudinal range was located above the respective lines of snow reliability. The same approach with varying baselines for natural snow reliability (i.e., lower in the Eastern part and higher in the Southern and South-Western part of the European Alps) was later applied in a much larger study for the OECD (Abegg et al. 2007). It was suggested to use the line of snow reliability to detect patterns in the geographical distribution of naturally snow-reliable ski areas. Snow conditions, however, are highly variable over space, and a highly aggregated line of snow reliability does not account for the complexity of the snow processes in mountain terrain (Steiger 2010). Further, it was recommended to not make statements on individual ski areas (Abegg et al. 2007: 32). The line of snow reliability, however, can be easily misinterpreted as a simple tool to rate individual ski areas, i.e., to distinguish between “winner” and “loser.” Financial institutions, for example, used the line of snow reliability to check on the ski areas’ credit ratings (e.g., Elsasser and Bürki 2002; Scott et al. 2007).
The ski season, as most tourism seasons, can be divided into different periods (e.g., Scott et al. 2007; Spandre et al. 2016a). Some periods are economically more important than others—an example being the Christmas holidays for mid-latitude ski areas in the Northern hemisphere. Consequently, Scott et al. (2008: 579) introduced the Christmas–New Year’s indicator as the “probability of being operational during the economically critical Christmas–New Year’s holiday period”, thereby referring to ≥ 30 cm of snow from 22 December to 2 January. Steiger and Abegg (2013) adopted the Christmas indicator, extended the period by 2 days (22 December to 4 January) and added the season opening indicator (snow depth ≥ 30 cm on 8 December). In Austria, the public holiday on 8 December (the Christian holy day of the Immaculate Conception)—or the weekend closest to this date—is the traditional opening day for many larger ski areas.
Many scholars use probability thresholds to rate the indicators. Wanner and Speck (1975) give an early example stating that their snow reliability criterion (≥ 90 days with ≥ 30 cm of snow) must be met in 90% of the winters. Often, several thresholds are available. It was suggested, for example, that the 100-day rule must be fulfilled in 70% (Abegg 1996) or 90% (Steiger and Mayer 2008) of the winters. Scott et al. (2008) suggested a 75% threshold for the Christmas–New Year’s indicator, and Steiger and Abegg (2013) who combined the 100-day rule, the Christmas indicator, and the season opening indicator used a 70% threshold for all three indicators. Steiger and Stötter (2013) went a step further and defined a “snow reliability classification scheme” considering the 100-day rule, the Christmas indicator, and the share of open skiing terrain. Probability thresholds (≥ 80%, 50–79%, and < 50%) were attributed to the Christmas indicator and the share of open skiing terrain resulting in a classification scheme ranging from “excellent” to “very poor.” There is a reasoning behind these thresholds: the indicators must not be fulfilled every year because of the changing quality of the snow season (“good versus bad snow years”) and the assumed ability of the ski areas to cope with a limited number of “bad” seasons and/or to compensate bad with good seasons. Why specific thresholds were chosen, however, often remains unclear. Some scholars (e.g., Abegg 1996; Steiger and Stötter 2013) double-checked with ski area operators, others did not—in general, information about stakeholder involvement is scarce.
Generally, snow days, season lengths, the 100-day rule, etc. can be calculated for both natural and technical snow. Snowmaking, however, shifts the perspective from natural conditions to more technical, infrastructural, and operational aspects. Scott et al. (2003) introduced a ski season and snowmaking simulation model (SkiSim). Snowmaking within SkiSim is based on technical capacities and operational practices which were derived from communication with ski industry stakeholders in Eastern Canada and include the start and the end date of snowmaking (22 November to 30 March), the snow base to maintain until 30 March (60 cm), the temperature required to start snowmaking (− 5 °C), and the snowmaking capacity (10 cm/day) (see also Scott et al. 2006, 2007). It is important to add that Scott et al. (2008: 583f) stated that “in order to compare the relative impact of projected climate change …, a single hypothetical ski area with identical characteristics (e.g., size, snowmaking capacities, and practices) was modelled at each study area. This approach isolates the importance of climate and projected climate change at each location, rather than assessing the relative technological (e.g., snowmaking) and business (e.g., four season operation) advantages of individual ski areas.” The SkiSim model was further developed (SkiSim 2.0) with refined snowmaking rules to better represent operational decisions over the ski season, i.e., distinction between base-layer snowmaking at the beginning of the season and reinforcement snowmaking later in the season (Steiger 2010). Applications in the European Alps with slightly adjusted snowmaking rules (e.g., snowmaking window from 1 November to 31 March instead of 22 November to 30 March) include Steiger and Abegg (2013) and Steiger and Stötter (2013). Pons et al. (2015) used a similar approach in the Pyrenees. Hanzer et al. (2014) developed a detailed, physically based model of technical snow production taking into account the ambient conditions and available snowmaking infrastructure. The model explicitly considers the topography and geometry of existing ski slopes, applies pre-defined snowmaking practices (i.e., base-layer and reinforcement snowmaking periods), incorporates grooming and the skier-induced downward transport of snow on the slopes, and tracks the water and energy consumption of the snow machines. In Marke et al. (2015), this model was applied to several ski areas in the Schladming region (Austria). Spandre et al. (2016b) implemented snow grooming and snowmaking in the physically based, multilayer snowpack model Crocus (Vionnet et al. 2012). This model was used to analyze past and future snow conditions in French ski resorts (Spandre et al. 2019a, b), both using the reliability line approach, and in an implementation where the spatial structure of each ski resort in a given region (French Alps) was explicitly represented. Hanzer et al. (2018) summarized the explicit implementation of snow grooming and snowmaking in the snow models AMUNDSEN (Strasser 2008; Hanzer et al. 2014; Strasser et al. 2019), Crocus (Vionnet et al. 2012; Spandre et al. 2016b), and SNOWPACK/Alpine3D (Bartelt and Lehning 2002). In summary, scholars have developed various, i.e., more or less standardized, ways to incorporate snowmaking into their snow models. It is interesting though how they deal with uncertainties. In most cases (at least in more recent studies), scholars take into account climate model ensembles and different emission scenarios but still rely on one single set of assumptions to define the snowmaking capacities and practices.
Other scholars focused on the snowmaking potential. Rixen et al. (2011), for example, calculated the number of potential snowmaking days, i.e., days with a dew point temperature ≤ − 4 °C; Hartl et al. (2018) did the same for German and Austrian weather stations, using a mean daily wet-bulb temperature threshold of − 2 °C, and computed the significance of historical trends. Hennessy et al. (2008) modelled potential snowmaking hours under current and future climate conditions in Australia, again using a − 2 °C wet-bulb temperature threshold. Similarly, Hendrikx and Hreinsson (2012) modelled potential snowmaking hours in New Zealand. They noted that a threshold of ≤ − 3 °C wet-bulb temperature is typically used in operational settings. However, they used a higher threshold of ≤ − 1.7 °C as discussions with the ski area operators indicated “that in warm years this was how the machines were operated, despite the cost and small amounts of snow that can be made at these warmer temperatures.” Spandre et al. (2015) analyzed past conditions of snowmaking hours in the French Alps and discussed the impact of various wet-bulb temperature thresholds.
Beyond the snow depth and duration indicators, additional climate variables play complementary roles in assessing the skiing conditions. Berghammer and Schmude (2014), for example, introduced the “Optimal Ski Day” as a day with no precipitation, a perceived temperature between − 5 and + 5 °C, more than 5 h of sunshine, and a wind speed less than 10 m/s, in addition to a minimum snow depth of 30 cm on the slopes and a white scenery in the surroundings. Demiroglu et al. (2018) identified the “ideal summer skiing day” through a consumer survey in Norway as one with no wind, an open sky, relatively warm temperature (10 to 20 °C), and wet or “corn” snow quality. Likewise, the China Meteorological Administration (2017) developed the “Meteorological Index of Skiing” by taking into account wind, temperature, and precipitation. Based on these antecedents and industrial inputs, a comprehensive climate index for ski tourism is proposed by Demiroglu et al. (2019). In this paper, however, we focus on the most essential component of ski tourism climatology—the snow cover—and do not provide an in-depth examination of individual or combined effects of the above-mentioned additional variables.