The present study was carried out in South Tyrol (Italy) and North Tyrol (Austria), both situated in the Central Alps. The landscape of the study site in South Tyrol, the Gsies valley, with a size of 109 km², is dominated by coniferous forests (50% of the municipal area), grassland with high land-use intensity (10% of the municipal area), and grassland with low land-use intensity in the subalpine–alpine zones (20% of the municipal area). Settlement takes up 5% of the valley floor. The municipality has experienced a big increase in forest (+12%) within the last 60 years and a decrease in grassland with low land-use intensity (-33%). The study site in North Tyrol, Neustift im Stubaital, with a size of 249 km², is characterized by a densely settled valley floor (1% of the municipal area is urbanized). The landscape is composed of rock and scree areas (33.6% of the municipal area), coniferous forests of the montane and subalpine zone (18%), alpine meadows and dwarf shrub communities (22%), and the Stubai Glacier (9%). In both study sites, arable farming has almost ceased.

To analyze the landscape features in relationship to distance and people's differences in the perception of scenic beauty, we used a standardized photograph-based questionnaire containing six photo series; each series was made up of four images: one real photograph and three different versions of the original photograph modified with Adobe PhotoShop^TM (Figure 1). There are different techniques for visualizing landscape scenarios, for example on-site visits, photographs, digital image editing and virtual landscape simulators [40]. For our study, we used digital image editing to generate natural-looking and photo-realistic images [19,41]. The technique allows changing only one variable in the picture, while the others are controlled or remain constant [40]. Simulated environments can be used to visualize hypothetical conditions [25,35].

To test the hypotheses 1 and 2, we used the three distance zones according to Schirpke et al. [31]: in the near zone (0–1.5 km), scattered elements (e.g., trees, row of trees, water, houses) and land-use types can be recognized; in the middle zone (1.5–10 km), elements and land-use types make up complexes of shapes and forms (e.g., different types of forest areas, diverse landscapes); in the far zone (>10 km), only mountain silhouettes with large-scale areas (forest, agricultural area, snowfield or glacier) are clearly identifiable. The different photo series were related to the hypotheses 1 and 2 as described in the following:

To answer hypothesis 3, the respondents were selected in public places on the basis of a stratified random sampling, with the aim to get more than 50 persons in each group (origin, gender, age, cultural background) (Table 1). The response rate was 89% and the 253 participants, locals and tourists of both study areas, were asked to rank the Series 1–6 by their scenic beauty without the possibility of equal rankings: from 1 = least beautiful to 4 = most beautiful. The respondents spent in average 5–10 minutes to fill out the questionnaire.

The rankings of the six different picture series were not comparable with each other because each series was related to a specific theme and the pictures were ranked only within each series. To compare the different themes, a seventh series was added at the end of the questionnaire, composed of the six original photographs from Series 1–6. The respondents were also asked to rank Series 7, ranking here ranging from 1 = least beautiful to 6 = most beautiful.

The statistical analyses were arranged in SPSS Statistics 17. The median of the ranked scenic beauty (total respondents) was calculated for all pictures. Each group (origin, gender, age, cultural background of tourists) was tested. We used the Mann-Whitney U test (p < 0.05) to compare the median values of the series. Based on Series 1, we derived weighting factors for the distance zones. By Series 2, the weighting result was verified. Series 3 to 6 were used to identify and quantify the influence of landscape elements by calculating landscape composition factors. Based on Series 7, a scenic beauty model was defined by using generalized linear models (GLM).

According to our hypotheses, depth of view and the closeness of specific landscape elements are important aspects of scenic beauty. To evaluate the influence of distance, but also the composition of the landscape in the near zone (positively or negatively perceived elements) on the perception process and, thus, on the scenic beauty of landscapes, we developed a general model allowing us to compare landscape properties and to evaluate its scenic beauty. We generated the scenic beauty model in two steps: (1) distance weighting factors, and (2) landscape composition factors.

Hypothesis 3 was not approved completely because there were significant differences in only 28.1% of all ratings in the different groups (cf. Table 3). In the gender group, only two significant differences could be found for pictures 2_1 and 6_3. The results of the origin group suggest that landscapes with human influence in the form of artificial elements (picture 5_1) are rated better by tourists than by locals. In addition, tourists also rated reforestation better than locals (pictures 2_1, 3_1). Analyzing the two cultural background groups of tourists (Italian, German), we found that Italian tourists rated artificial elements (picture 4_3) and reforestation (picture 2_1) in the landscape better than German tourists did. The age group returned particularly significant differences between respondents over 60 years and those under 60 years. The older generation rated open, structured landscapes better than younger respondents did (pictures 5_1, 6_1, 3_4). As hypothesis 3 was not verified, we did not include specific perception of different socio-demographic groups in our modeling approach.

The results suggest that distance played a major role in the perception of Alpine landscapes. Topography influences the depth of view which is central to our understanding of landscapes and perceived scenic beauty [32,34]. As confirmed by Tveit [8], the distance from the observer was important for the perception of scenic beauty because people’s ability to distinguish between features is decreasing with distance [47]. We used the distance zones, adapted to the Alps, by Schirpke et al. [31]: the near zone 0–1.5 km, the middle zone 1.5–10 km, and the far zone 10–50 km. For each distance zone, we derived the distance weighting factors: near zone 64%, middle zone 22%, and far zone 14%.

Landscape attributes and elements like naturalness, diversity and urban sprawl are often used as predictors for scenic beauty, however, mostly not related to view zones [11,21,24,25,26]. Similar to other studies [21,34,43,44,45], attributes like naturalness, presence of water, and diversity of elements emerged as positive predictors for scenic beauty in the near zone. In contrast, the modern human settlement in the near zone and also human influence as ski tourism infrastructure in the middle zone were evaluated negatively. An unexpected result emerged for the landscape diversity of the middle zone (Series 4). In the literature, diversity is generally considered an attribute for a high scenic beauty of landscapes [28,41,46]. Accordingly, Lindemann-Matthies et al. [41] found a high preference for diverse landscapes with structural elements in the middle zone but a low rating for arable land. In our study, it was not the most diverse landscape that was rated highest but a medium structured grassland [48]. This result may be explained by the view of the valley seen from above and at a larger distance. Furthermore, the results indicate that the younger generation (under 60 years) was not familiar with the pattern of a diverse landscape and that for the middle zone concepts like coherence [1,20,44] and legibility [26] were more important than diversity [20,26,44].

Contrary to the hypothesis 3, no significant differences emerged for 71.9% of all ratings in the different groups indicating that the landscape variations were much greater than the variations of observer’s judgments as also confirmed by other studies [29,36,37,38]. Although 28.1% of all ratings were significantly different, the median values are in most cases the same. Our study therefore reflects the perceptions of all respondents without distinguishing between different socio-demographic groups [13]. Nevertheless, some differences have to be pointed out. Locals in South and North Tyrol evaluated landscape changes more critical than tourists. The local group seemed to be attached to the current landscape or would prefer a development towards traditional cultural landscapes. In contrast, the tourist group seemed to be more open to greater landscape changes like reforestation or human-caused landscape elements.

According to our results, the locals are familiar with the cultural landscape through their professional or leisure uses of it [27]. Big changes (e.g., reforestation or urban sprawl) may be interpreted as loss of home land [4]. Unlike younger people, the 60+ generation also preferred traditional cultural landscapes with great landscape diversity. With the decrease of diverse landscapes today, people are not familiar any more with landscape diversity and thus not able to read the landscapes [26]. The older generation, however, has known these landscapes from childhood. As the perception of scenic beauty starts at an early age, the older generation rated this landscape type better [35].

Tourists rated a landscape characterized by artificial elements and facilities but also natural reforestation higher than locals did. In tourist destinations like South Tyrol and Tyrol, skiing areas and the related facilities offer many holiday options for tourists. Hunziker [4] stated that an ideal landscape depends not only on its scenic beauty but also on other interests. If a ski lift is relevant for touristic interests, then the tourists will rate such a landscape more positively. Moreover, differences between the German- and Italian-speaking groups could be found. While the Italians were more open to extreme landscape changes, for the Germans among the tourist respondents, naturalness was more important.

Although the questionnaire consisted of a small number of pictures and general conclusions are limited, our results indicate that the modeling approach is a suitable option for creating a scenic beauty model. To validate the model, the study should be repeated with a higher number of pictures indicating clearly the different distance zones and testing various landscape elements especially in the near zone. Furthermore, the ranking of the pictures was problematic for comparing the different series among each other. The pictures should therefore be rated to allow a direct comparison.