A landscape menu to please them all: Relating users’ preferences to land cover classes in the Mediterranean region of Alentejo, Southern Portugal

Land Use Policy 54 (2016) 355–365

Contents lists available at ScienceDirect

Land Use Policy journal homepage: www.elsevier.com/locate/landusepol

A landscape menu to please them all: Relating users’ preferences to land cover classes in the Mediterranean region of Alentejo, Southern Portugal Diana Surová a,∗ , Teresa Pinto-Correia a,b a

ICAAM— Instituto de Ciências Agrárias e Ambientais Mediterrânicas, Universidade de Évora, Núcleo da Mitra, Ap. 94, 7002-554 Évora, Portugal Departamento de Paisagem, Ambiente e Ordenamento, Escola de Ciências e Tecnologia, Universidade de Évora, Colégio Luis António Verney, Rua Romão Ramalho, 59, 7000-671 Évora, Portugal b

a r t i c l e

i n f o

Article history: Received 26 July 2015 Received in revised form 30 January 2016 Accepted 22 February 2016 Keywords: User groups Values Landscape preferences Services Spatial planning Complex land use systems

a b s t r a c t To address multifunctional land-use management, and as a response to societal expectations with regard to agriculture and well-being in rural spaces, one of the aspects in need of particular attention in the current research is the spatial expression of the landscape attractiveness. Current knowledge about the attributes of landscape that are most highly appreciated, and the ways that individual differences may influence landscape preferences, is still incomplete, particularly for complex agricultural landscapes in the Mediterranean. The aim of this paper is to provide an assessment, including quantitative and qualitative approaches, of user-based landscape preferences in the Mediterranean region of Alentejo, Southern Portugal. To allow for the spatial expression of these preferences and the evaluation of the impact of changes to land-use, specific land cover classes have been used as the basis for our survey. The results are aimed at adding to the body of research literature on the diversity of the landscape attributes that are valued by different user groups and at contributing to informed decision-making processes when resolving land-use issues in the region being studied. To attain this objective, a photo-questionnaire was conducted to identify preferred land cover classes and the land cover qualities that landscape users appreciate most. The study shows that the way that landscape is used is a significant factor influencing the preferences, following a clear functional demand pattern. Not only are there differences between land managers as producers and others as consumers, but the appreciated land cover qualities and land cover preferences also differ from one consumer group to another. Knowledge regarding the preferences of specific groups can inform landscape management at different levels of governance in such a way that multi-functionality may be more successfully attained. Moreover, certain particularities with regard to the preferred landscapes in the region being studied are discussed in comparison with the results of earlier studies from other regions. In this regard, this paper puts great weight on the importance of territorial contextualization in regard to making general assessments concerning landscape preferences. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Landscape quality objectives play an important role in several European policies, particularly in the Common Agricultural Policy (CAP) and at various levels of the Structural Fund programs (Dax, 2014), aimed at attaining sustainable rural development, territo-

∗ *Corresponding author. E-mail addresses: [email protected], [email protected] (D. Surová), [email protected] (T. Pinto-Correia). http://dx.doi.org/10.1016/j.landusepol.2016.02.026 0264-8377/© 2016 Elsevier Ltd. All rights reserved.

rial cohesion and the conservation of natural resources (Johansson et al., 2012; Jones and Stenseke, 2011; European Commission, 2010). In addition, the European Landscape Convention (ELC) encourages public authorities to adopt policies and measures aimed at protecting, managing and planning landscapes to maintain and improve their quality and level of recognition (Council of Europe, 2000). With the current rapid and profound changes in rural areas, assigned mainly to changing levels of agricultural land-use intensity, to increasingly urbanised society and to life-style changes (Primdahl and Swaffield, 2010; Primdahl et al., 2013), rural land-

356

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

scapes need effective management interventions that will maintain or improve delivery of their goods and services. Agriculture, as a main land use in rural areas, is the principal sculptor of landscape, and it is expected to satisfy societal demands in a way to provide multi-functional spaces that fulfil both productive as well as non-productive functions (Holmes, 2012). At the same time, agricultural management in rural areas plays an important role in issues of food security, shortages in natural resources, and climate change. These matters lead to increasing interest and effort of policy and research in attainment of sustainable rural development focusing on territorial cohesion, conservation of natural resources and rural well-being. Consequently, the scientific community is facing the challenge of producing reliable, comprehensible, and “solution-oriented” data to support the drafting of effective policy regulations. With regard to land-use management, generating data calls for a shift from a sector-based to a geographical perspective, given that most of the functions that society expects from rural areas vary from one geographical landscape to another (Selman, 2012). Moreover, the involvement of non-academic stakeholders in academic research is crucial (Marsden and Sonnino, 2008), in such a way that all involved user groups are considered, to effectively contribute to the creation of a new science-policy-society dialog (Pohl, 2008). The need to recognize the diversity of society’s preference to landscape use is upheld in the ELC as well as in its Recommendations for Implementation (CM, 2008). These documents stress the importance of identifying the link between society’s requirements and the values the public attaches to different landscapes. Although a wide range of landscape indicators is often used in monitoring processes, the public’s evaluation of landscape is often disregarded (Paracchini et al., 2015; Pinto-Correia and Kristensen, 2013), even if some of the services, such as cultural, rely fully on public recognition. The need for sound scientific preference data as a basis for generating operational indicators has been an ongoing demand in recent years (Paracchini et al., 2015; Pinto-Correia et al., 2014). In particular, urgent research requirements have been the indicators of users’ landscape preferences, adapted to different regional landscape typologies, (Jones et al., 2015; Paracchini et al., 2012; Pinto-Correia and Carvalho-Ribeiro, 2012), which also involves the spatial expression of landscape attractiveness (Carvalho-Ribeiro et al., 2016; Paracchini et al., 2012). The objective mapping, monitoring and quantification of land-user preferences requires preference data regarding existing mapped units, which allow for the upscaling of locally collected data (Jones et al., 2015). The Mediterranean areas are unique in Europe in terms of the predominance of silvo-pastoral systems, in the dynamics of changing agricultural processes (Pinto-Correia et al., 2011), and also at the level of land-use systems in the absence of land units’ clear boundaries (Barroso et al., 2012). The silvo-pastoral land-use systems are subject to agricultural competitiveness but also provide a wide range of goods and services and may therefore be considered highly multifunctional. However, to date, there has been a surprising lack of research exploring these landscapes from the perspective of different user groups (Pinto-Correia et al., 2014). The aim of this paper is to assess user-based preferences for land cover classes in the rural Mediterranean region of Alentejo, Southern Portugal, where extensive traditional land-use systems are still maintained. The results aim to contribute to the research literature about appreciated landscape characteristics of different user groups as well as to contribute to informed decision-making processes in land-use planning issues and the drafting of policy in the region being studied. To attain this goal, a photo-questionnaire has been conducted in the region being examined to ascertain those factors influencing landscape preferences and expected qualities. The survey uses land cover classes as components of these landscapes

to facilitate the transferal of the results onto spatial analyses and scenarios (Carvalho-Ribeiro et al., 2013). This paper consists of the following sections. Following this Introduction, there is a short literature review of papers that address the evaluation of landscape preferences among different user groups. The next section explains the methodology applied, followed by the presentation of the findings. Following that, there is a discussion of the results and methodology, and the regional particularities together with some implications of the findings for policy makers are addressed; finally, our conclusions are outlined.

2. Literature review Research on landscape preferences shows that there is no consensus between all members of the public (Tudor et al., 2014; Howley, 2011; Buijs et al., 2006). The way people evaluate landscape depends on a variety of factors, such as socio-demographic characteristics (e.g., Junge et al., 2011), cultural and economic aspects, and the individual’s values and ways of interacting with the landscape. Some studies have assessed landscape preferences from perspectives related to different user groups. Purcell (2006) showed that, among other characteristics, nationality as well as a sort of preference judgment with regard to where one lives, works or goes on holiday could inform one’s preferences. More recently, Hofmann et al. (2012) have illustrated the differences between residents and landscape planners with regard to preferences for urban parks. Farmers and non-farmers had different opinions when it came to assessing the attractiveness of Swiss landscapes (Junge et al., 2011). The divergences in landscape appreciation between farmers and naturalists (Natori and Chenoweth, 2008); among farmers, landscape experts and country-dwellers (Rogge et al., 2007); and also between residents and tourists (Rambonilaza and DacharyBernard, 2007) were demonstrated. Searching exclusively for the preferences of one general group of landscape users may bias the results (e.g., Berninger et al., 2010), thereby running the risk of mismatches in policy targeting. It is as such that the multi-functionality of landscapes may be more deeply grasped when the public is divided into different user groups and their specific perspectives, values and requirements are assessed (Pinto-Correia et al., 2014). Over the years of research conducted into landscape preferences, no standardized assessment method has been established. As such, it is possible to identify a number of different approaches in published studies (Van Zanten et al., 2014). Preference data may be gathered by survey or through expert knowledge. Surveys of the general public are generally considered to generate more accurate information regarding societal values (Tveit, 2009). The two main types of surveys are large-scale quantitative surveys of the public and qualitative, normally small-scale surveys. The first type has the advantage of providing statistically valid results. However, qualitative methods allow for more exploration of the respondents’ points of view, complex ideas, values and expectations and are highly regarded for generating a wealth of information. The multi-method strategy that combines both qualitative and quantitative surveys in an integrative manner delivers ‘the best of both worlds’ in terms of their capacity for generalized explanation (Davis and Michelle, 2011). The combination of both can help to build partial and complementary pictures of as complex a reality as that of the public’s landscape preferences. Considering the above mentioned statement, and with an intention to choose the best possible approach for the issue at hand, the present study combines quantitative, close-ended questions with qualitative, open-ended questions in a large-scale survey applied to the wide public of landscape users in the studied area.

Table 1 The socio-demographic characteristics of the respondents.

Socio-demographic attributes

User groups in percentages

N and percentage

Inhabitants

New rural inhabitants

Second-home residents

Regular visitors

Tourists

Eco-tourists

Hunters

Gender

Women/men

330/736 31/69 48/52

50/50

49/51

30/70

55/45

43/57

Nationality

Portuguese/other

893/173 84/16 98/2

53/47

93/7

85/15

51/49

75/25

University degree

No/yes

672/394 63/37 68/32

38/62

46/54

50/50

40/60

61/49

90/10

75/25

Age class

<=40/>40

360/706 34/66 39/61

38/62

28/72

44/56

45/55

54/46

25/75

16/84

Childhood/current residence place

Alentejo Other Portuguese place Out of Portugal

581/756 55/71 89/100 310/219 29/20 10/0 175/91 16/8 2/0

6/100 50/0 44/0

38/16 55/73 7/11

39/41 44/51 16/8

7/6 44/49 49/45

45/56 31/22 24/22

87/91 11/9 1/0

79/96 16/4 5/1

Residence background

Urban Urban-to-rural Rural-to-urban Rural

157 151 108 650

30

45 7 35 14

29 8 21 42

56 9 19 16

18 9 16 57

2 3 5 90

3 9 1 87

Farming background

Yes/no

653/413 61/39 53/47

61/39

46/54

52/48

34/66

32/68

69/31

Professional status

Employed Unemployed Student Retired

803 37 29 197

75 3 3 18

73 3 4 21

84 5 2 10

62 1 1 35

73 5 4 18

63 5 11 22

82 5 3 11

83 3 0 14

73 2 1 24

Profession

Agriculture, fishing, forestry Construction and industry Services and private sector

280 102 684

26 10 64

20 7 73

14 4 82

1 11 88

7 15 78

0 6 94

3 6 91

27 19 54

83 6 11

100

11 114

12 133

7 74

11 117

10 104

11 114

19 204

19 206

Respondents (%) N

1066

15 14 10 61

4 96

70

1/99

Land managers

100/0

14/86 98/2

100/0

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

Total

357

358

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

Table 2 Examples of responses to open-ended question about perceived qualities of preferred land-covers, organized in six categories. Aesthetic I like its organization/the visual structure Because of the color contrast between brown soil green vegetation and the blue sky I like the cork oak tree; it is beautiful Cereal seems like a sea shaped by the wind I like these shrubs; they smell good It changes colors during the year

Identity It is the Alentejo It is different from other places I am from the Alentejo; i must like it! It is typical and traditional It mediates my grandfather’s and father’s lives I have grown up in this landscape

Production It provides an income It contributes to regional development It needs human work and brings employment It reduces food imports All of us need it to have food Provides basic needs for a nation The land needs to produce

Wilderness It is vast, disordered, and spontaneous There is an absence of human intervention This one is the most natural and uncultivated There is abundant vegetation; it is dense with much green It is close to the natural state It is wild

Ecology It is well adapted to local climate conditions Forests produce oxygen It protects the soil There can be found a variety of animal species It provides shelter for wild species

Recreation There is an open field convenient for hunting Here the hunting is more difficult and thus more interesting This is good for hiking (strolling) This has clean ground; it is good for picnics I like to go there and taste the local vine

3. Methods 3.1. Study area The study area was the Alentejo NUT II region in Southern Portugal. The population density of this region has been continuously falling over recent decades, and in 2013, it stood at 23.5 people per square kilometer, while the national population density was 113.1 people per square kilometer (INE, 2014). The region has a typical Mediterranean climate with very hot and dry summers and most rain falling during the winter season, where in particular, the silvo-pastoral “montado” predominates. The landscape in this region is highly prized for a variety of leisure and recreational activities that are not considered in statistical databases. The municipalities where the questionnaire was conducted were chosen according to two consecutive approaches: automatic and expert-based. The first approach, more automatic and quantitative, was developed through a classification of the Corine Land Cover (CLC) distribution in 2006 across the municipalities. In the groups, municipalities with similar CLC distribution were assembled. Subsequently, an expert panel, including members from both research teams and members of the regional development agencies, selected the case-study municipalities. The aim of this process of municipalities’ selection was to obtain a representation of the region’s diversity of land cover patterns and socio-economic dynamics. Accordingly, the following municipalities were selected: Castelo de Vide, Ponte de Sor, Elvas, Montemor-o-Novo, Reguengos de Monsaraz, Grândola, Ferreira do Alentejo, Vidigueira, Serpa, and Almodôvar. 3.2. Sample design The questionnaire was devised to survey a diversified sample of landscape user groups. The main focus was on existing landscape consumption uses in the region and on land management. There was information used from previous research (Surová and PintoCorreia, 2008) to identify classes of potential users of landscape (e.g., for hunting), and this information was used to stratify the sample. Accordingly, the selection included long-term inhabitants, new rural inhabitants, second-home residents, frequent visitors, multipurpose tourists, eco-tourists, hunters, and land managers. People who fell into these classes, in approximately equal number in each class, were recruited.

Long-term inhabitants were people who had lived permanently in the case study municipalities since their childhood. New rural inhabitants were those had had come to settle in the region during their adulthood. Second-home residents owned a house or an apartment in which they spent weekends or holiday time. Those visiting a surveyed municipality several times per year without having a house there were labelled as regular visitors. Multipurpose tourists were people enjoying the landscape in the study area by car or on organized bus-tours, focusing on different aspects such as heritage, cuisine, and cultural events, etc. However, the group labelled as eco-tourists enjoyed the landscape mainly through outdoor activities such as hiking and biking. Hunters were people undertaking hunting activities in the region. Land managers were those making decisions regarding agricultural or forest activities on private properties in the study area. To begin with, contact was established with municipality offices and hunting, tourist, and land managers’ associations in the study areas to make initial contact with potential respondents. This was followed by snowball sampling, in the interests of enhancing the representative nature of the sample groups and to select the people who were best suited to the requirements of the survey. The survey was then conducted with those who volunteered to take part. 3.3. Photo questionnaire A quantitative approach, using a photo questionnaire, was used to identify the respondents’ preferences for CLC classes. A qualitative selection, for which open-ended questions were used (Patton, 2002), was also applied in the interest of getting a clearer picture of the qualities that respondents attributed to land covers in a specific context. The questionnaire was conducted in a face-to-face manner. At the beginning of each questionnaire, the respondent was given a brief introduction to the purposes of the study. The survey was conducted in Portuguese, and for those with scant knowledge of the local language, English was also available. First, each respondent was designated to one of the predefined user groups, according to his/her main activity in the study area. Although many of the respondents could in fact be designated to more than one user group, for example, hunters and habitants, they were asked to choose just one and to respond as such during the entire questionnaire. Subsequently, they chose the three CLC classes they liked the most from a set consisting of 16 photos. Then, they were asked to explain their choices through open-ended questions. In the last

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

part of the inquiry, information was collated regarding the sociodemographic background of the respondents. The photographs, used as visual stimuli, consisted of 15 different CLC classes currently dominating the Alentejo landscape. An additional 16th scene was included, representing intensive olive farming, recently spreading in the region to replace traditional olive groves. Intensive and super-intensive olive groves differ substantially in appearance from traditional olive groves (Beaufoy, 2001). Each photo depicted one single CLC class. Land cover data are recognized as being a useful basis for landscape evaluation, due to the spatial applicability and increasing availability of datasets. It has the practical advantage of being directly observable in the field as well as from remote devices, for example, through satellite images. The photographs taken in the field were edited using the Adobe Photoshop CS3 graphic editing program to eliminate any elements that might distract from the focus of the survey. These included human artefacts such as roads, electric pylons, and walls, etc. Furthermore, the same sky and the same horizon level were applied to each photo. All 16 photographs used in this survey were presented in the publication of Carvalho-Ribeiro et al. (2013). 3.4. Data analysis Multivariate analyses were applied to collected data. In comparison with univariate analyses, these analyses facilitate the identification of relationships among preferred CLC classes, perceived land cover qualities and the characteristics of land users. The content analysis was applied to analyze qualitative data concerning the perceived land cover qualities of chosen CLCs. The full responses to the open-ended questions were codified and grouped into a reduced number of classes representing different meanings, according to the content analysis process (Patton, 2002). A team-based qualitative analysis strategy was used to develop the list of categories. First, a theoretical basis from Buijs et al. (2006) and Swanwick (2009) was used to organize responses according to different meanings. Afterwards, through several iterations, members of the research team determined categories and subcategories of responses to open-ended questions. The research team subsequently reviewed the coding structure, resolved discrepancies, and reviewed the responses. The process continued until total consensus between research team members was achieved for the final list of categories. The responses were then codified for subsequent analysis. Because most of the data were nominal, multiple correspondence analysis (MCA) with subsequent Cluster analysis and finally Multinomial logistic regression (MLR) were applied. The software used for analyses was SPSS 18.0 (manufacturer: IBM 214 Corporation, New York, USA). The MCA allows for the simultaneous analysis of more than two categorical variables (Greenacre, 2006) and provides the identification of relatively homogeneous groups. MCA submits qualitative data to the processes of quantification. The object scores from MCA were used for Hierarchical and K-means cluster analysis to identify similarities between preferred land covers, perceived qualities and user-group membership. In hierarchical cluster analysis, the agglomeration schedule statistic with Ward’s method was applied, and the Squared Euclidean distance was chosen. The scree plot of coefficients from the agglomeration schedule was then analyzed, and the step of ‘elbow’ identified to determine the number of clusters. Afterwards, the K-means cluster analysis was applied, and the cluster membership of each object was saved as a new variable. MLR, as a predictive analysis, was used to model the relationship between cluster membership and socio-demographic characteristics of respondents. The dependent variable in MLR was a discrete choice between identified clusters. Cross tabulations between each of the categorical predictors with response variable were applied,

359

and the multicollinearity of predictors checked. Next, the forward stepwise independent variable selection was applied to all eligible variables, and the final model was calculated. Additionally, the landscape cluster probability membership in percentage per predictor level was calculated according to individuals’ realized predictor values. 4. Results The survey totalled 1,066 fully completed questionnaires over the course of eight months. In each of the ten study municipalities, approximately 100 face-to-face questionnaires were applied. Respondents from all predefined user groups were surveyed in each municipality. Table 1 depicts the socio-demographic characteristics of the respondents in the sample. Considering the entire sample of respondents, the most preferred CLC class was montado (52%), followed by an agricultural mosaic (38.6%) and non-intensive olive groves (37.7%). The least preferred land cover class was eucalyptus groves (3.6%), followed by irrigated pasture (4.6%) and rice fields (5.3%). Before progressing to further analyses, the responses to the open-ended question were categorized and turned into quantitative variables. 4.1. Qualitative data results We established the following six categories of land cover qualities that were perceived as being expressed by the preferred land cover classes: aesthetics, identity, production, wilderness, ecology, and recreation. The aesthetic category was related to responses expressing an appreciation of visual aspects, such as colors and shapes. Cultural and historical qualities were grouped along with the landscape’s uniqueness in the identity category. Meanwhile, the production category related to functional representations such as food production or the satisfaction of primary life requirements. The wilderness category included the appreciation of aspects perceived as representing nature unaffected by the presence of humankind. The ecological category involved indirect benefits to society such as biodiversity and environmental and climate regulation. Finally, the recreational category expressed the favorable physical conditions of a particular land cover in terms of leisure activities. Examples of responses for the six land cover qualities are displayed in Table 2. The frequency with which the six perceived quality categories were mentioned, the preferences for land cover classes expressed, and the Chi-square test values are shown in Table 3. The most frequently mentioned land cover quality was identity, which was particularly attributed to non-intensive olive groves, montado and cereal crops. The second most frequently mentioned quality was the aesthetic one, particularly in relation to agricultural mosaics and pine forests. The wilderness quality was significantly more frequently mentioned for forest areas. The quality that was least frequently mentioned was the recreational one, with reference to land cover providing favorable physical conditions for outdoor activities. 4.2. Clustering of land cover preferences and perceived land cover qualities MCA with a variable principal normalization method was applied to data on preferred land cover types and their perceived qualities. The initial MCA calculation included responses to all land covers and to all variables of perceived qualities. The model resulted in ten dimensions with Eigenvalues > 1 explaining 61.8% of the variance. To improve the model, the variables with discrimination measures <0.2 in the first three dimensions were omitted from the subsequent MCA. The final MCA included the following categorical land cover preference variables: irrigation crops, montado

360

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

Table 3 The count frequencies and expected frequencies in parentheses of six land cover quality categories, preferences for land cover classes, and the Chi-square test values with significant levels of differences between user groups. User groups INH

NRI

SER

RVI

TOU

ECT

HUN

LM

Total

PCS

Perceived qualities Aesthetic Identity Production Wilderness Ecology Recreation

38 (32) 64 (50) 50 (41) 14 (17) 18 (30) 26 (21)

56 (38) 69 (59) 45 (48) 38 (20) 23 (35) 21 (25)

22 (21) 43 (33) 25 (27) 12 (11) 7 (19) 12 (14)

38 (33) 63 (52) 49 (42) 17 (18) 15 (31) 17 (22)

50 (30) 38 (46) 26 (38) 28 (16) 7 (27) 13 (20)

48 (32) 52 (50) 17 (41) 34 (17) 12 (30) 28 (21)

16 (58) 30 (90) 25 (74) 5 (31) 164 (53) 51 (38)

35 (59) 111 (91) 148 (74) 12 (31) 32 (54) 32 (39)

303 470 385 160 278 200

100.5*** 102.4*** 197.2*** 90.0*** 391.4*** 15.5*

Land covers CR IC RF VN OR OG IP MS MN EC PN MF PS HS LS IO

30 (34) 35 (20) 7 (6) 50 (31) 10 (9) 37 (43) 5 (5) 45 (44) 47 (59) 5 (4) 18 (14) 21 (24) 6 (11) 6 (13) 5 (15) 14 (9)

32 (40) 17 (23) 7 (7) 34 (36) 12 (10) 48 (50) 3 (6) 80 (51) 78 (69) 3 (5) 11 (17) 40 (28) 4 (12) 17 (15) 7 (17) 5 (10)

25 (22) 10 (13) 6 (4) 25 (20) 7 (6) 34 (28) 2 (3) 34 (27) 37 (38) 3 (3) 7 (9) 17 (15) 4 (7) 3 (9) 4 (10) 4 (6)

36 (35) 27 (20) 5 (6) 38 (32) 8 (9) 43 (44) 4 (5) 50 (45) 57 (61) 0 (4) 18 (15) 23 (24) 1 (11) 16 (14) 10 (15) 15 (9)

16 (31) 12 (18) 7 (6) 25 (29) 9 (8) 41 (39) 3 (5) 46 (40) 54 (54) 8 (4) 25 (13) 29 (22) 6 (10) 12 (12) 12 (14) 6 (8)

29 (34) 12 (20) 6 (6) 23 (31) 9 (9) 39 (43) 3 (5) 44 (44) 65 (59) 6 (4) 21 (14) 39 (24) 6 (11) 24 (13) 13 (15) 3 (9)

73 (61) 12 (35) 10 (11) 25 (56) 5 (16) 89 (77) 6 (9) 45 (79) 98 (106) 12 (7) 21 (26) 36 (42) 54 (19) 39 (24) 80 (27) 6 (16)

80 (62) 58 (35) 9 (11) 72 (56) 24 (16) 71 (78) 23 (9) 68 (80) 118 (107) 1 (7) 13 (26) 17 (43) 19 (19) 6 (24) 8 (27) 31 (16)

321 183 57 292 84 402 49 412 554 38 134 222 100 123 139 84

26.1*** 61.7*** NS 51.8*** NS NS 25.9** 56.3*** NS 20.2** 29.1*** 44.1*** 94.0*** 45.8*** 159.5*** 37.1***

NS— not significant. Abbreviations of User groups: INH—inhabitants; NRI—new rural inhabitants; SER—second-home residents; RVI—regular visitors; TOU—tourists; ECT—ecotourists; HUN—hunters; LM—land managers. Abbreviations of Land Covers: CR—cereal crops; IC—irrigation crops; IO—intensive olive grove; VN—vineyard; PN—pine forest; MN—montado; OG—olive grove; MS—agricultural mosaic; MF—mixed forest; PS—natural pasture; LS—low shrubs; HS—high shrubs in forest land. * Significant at the 0.05 level. ** Significant at the 0.01 level. *** Significant at the 0.001 level. Table 4 The model summary resulting from the final multiple correspondence analysis.

WIL

Model summary Variance accounted for Inertia % of variance Total (Eigenvalue)

1 2 3 Total Mean

1.717 1.387 1.257 4.362 1.454

a

.364a

.245 .198 .180 .623 .208

24.536 19.819 17.954 62.308 20.769

CR IC

80

AES

Dimension Cronbach’s Alpha .487 .326 .238

100

IO

60 PRO

VN

40 20

PN

IDE

0

Mean Cronbach’s Alpha is based on the mean Eigenvalue.

MN

ECO

and low shrubs; and the following perceived quality variables: aesthetics, identity, production, and ecology. In this model, the first three dimensions explained 62.31% of the variance (Table 4). The object scores from the three dimensions were saved as the new variables used for the subsequent Hierarchical cluster analyses. As the scree plot from the agglomeration schedule shows, the differences between the coefficient values of within-cluster sum of squares in relation to the number of clusters start to become relatively insignificant once four or more clusters are reached. As such, the number of four clusters was shown to be the best option. The subsequent K-means cluster analysis divided the sample into four different landscape clusters based on the object scores from the final MCA. The principal distinctive characteristics of the clusters identified are shown in Chart 1. The first cluster, titled open natural landscape (n = 228), was specific in terms of its perceived ecological quality, with high preference levels expressed for low shrubs on agricultural land and preference expressed fairly frequently for natural pasture and high shrubs on forest land. In this cluster, there was practically zero preference for intensive agriculture. The second cluster, identity landscape (n = 396), was characterized by the perceived qualities pertaining to identity, and showed a

OG

HS MS

LS PS

Open natural

Identity

MF

Production

Aesthetic

Chart 1. The main characteristics of the four landscape clusters. Abbreviations: CR—cereal crops; IC—irrigation crops; IO—intensive olive grove; VN—vineyard; PN—pine forest; MN—montado; OG—olive grove; MS—agricultural mosaic; MF—mixed forest; PS—natural pasture; LS—low shrubs; HS—high shrubs in forest land; perceived qualities: ECO—ecological; IDE—identity; PRO—production; AES—aesthetic; WIL—wilderness.

high level of preference for montado, non-intensive olive groves and cereal crops. The next cluster, production landscape (n = 190), stood out largely by its land cover types featuring perceived production qualities such as intensive agriculture. Within this cluster, irrigation crops were the most frequently preferred land cover type. The fourth cluster, aesthetic landscape (n = 252), may be recognized mainly by its aesthetic qualities as well as by its aspects

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

of wilderness. These were represented most commonly by agricultural mosaics, although also on occasion by montado. Pine and mixed forests were also popular in this cluster. 4.3. Landscape cluster preference predictors To identify the factors influencing people’s preferences for specific landscape clusters, we applied MLR. The response variable presented four outcomes corresponding to four clusters. The following socio-demographic characteristics were taken as predictor variables: nationality, urban–rural background, user-group, childhood and current residence place, relation to agriculture, education, principal employment activity, profession, gender, and age range. Each time a categorical predictor was cross-tabulated with a response variable, the result was significant with the Pearson chi-square statistic (<0.05), other than in the profession and principal activity variables. Variables such as nationality, urban–rural characteristics, current place of residence and childhood residence showed significant correlations. Consequently, only one variable was selected for MLA, the childhood place of residence. The forward stepwise independent variable selection was applied to six eligible variables: user-group, childhood place of residence, relation to agriculture, education, gender and age. The analysis resulted in three significant predictor variables: user-group, education and gender. A likelihood ratio test showed that the model fit the data better than the null model did. The final model was 55.5% accurate. The parameter estimates (Table 5) shows the logistic coefficient (B) for each predictor variable for each alternative landscape cluster category. The open natural landscape and hunter user group were used as reference categories. The landscape cluster probability membership in percentage per predictor level is shown in Table 6. This showed, for instance, that new rural inhabitants are most likely to prefer identity landscape (43.6%) and least likely to prefer production landscape (10.5%). In the case of tourists visiting the Alentejo region, it is most likely that they will prefer aesthetic landscape (47.1%), while it is not likely that they will like production landscape (10.6%) or open natural landscape (9.6%). People with university degrees are most likely to prefer aesthetic landscape (22.5%) as opposed to production landscape (14.4%). However, in the case of men, there is no considerable difference between preferring production landscape (17.5%) and wilderness landscape (18.2%). Landscape users who had spent their childhood outside Portugal were most likely to prefer wilderness landscapes (38.7%). Furthermore, people without any relation to agriculture were less likely to prefer production landscapes (13.8%). Examining user groups, it became clear that some of the landscape clusters have relatively clear profiles. People who prefer open natural landscape in the Alentejo region are mostly hunters (77.5%), and they are rarely women (8.5%). People inclined toward production landscape were frequently land managers (34.5%). Additionally, tourists (47.1%), eco-tourists (39.5%) and new rural inhabitants (34.5%) are much more likely to prefer aesthetic landscape to any other landscape. However, people favoring identity landscapes are more diverse in terms of user group, but it is highly probable for them to have university degrees (49.5%). 5. Discussion The land cover preferences are dependent on the way the landscape is used. Different user groups look for different types of land cover and appreciate different land cover qualities. Earlier studies, even if not in the Mediterranean region, have also observed these trends (Purcell, 2006). Much like in certain previous studies from other regions (e.g., Burton et al., 2008; Buijs et al., 2006), land managers express a

361

markedly greater preference for production landscapes than other users do. In the Alentejo, the production-based perspective is currently most closely associated with preferences for irrigation crops, vineyards, intensive olive groves and orchards. However, considering all of the land managers interviewed, a substantial proportion of them would place a positive value on identity landscape cluster, reflecting a tie with previous generations and the sense of family heritage. This study did not distinguish between the preferences of different land manager types such as, for example, small-scale resident farmers and land managers with large properties having less family-related ties with their land. As recent research shows, land manager typology may be a relevant issue for understanding their values and thus the potential management decisions and landscape trajectories of the future (Giannoccaro and Berbel, 2013). The hunter perspective has its own unique take on the landscape, differing from both landscape consumer users and farmers. Hunters are much more likely to look for open natural areas, as these areas provide suitable habitats for game animals. Moreover, these land covers without trees provide good visibility and security to hunters during the hunt (Surová and Pinto-Correia, 2008). Hunters do not seem to be principally concerned about aesthetic, wilderness aspects or identity land cover qualities. In Alentejo, the general aesthetic qualities of the land covers, such as visual structure, color contrasts, a perceived sense of wilderness or an agricultural mosaic, are appreciated particularly by passing visitors, such as tourists, or by those who did not spend their childhood in the area. This result may indicate that the level of familiarity with the region can influence landscape preferences. Other researchers have also considered a level of familiarity with the landscape or respondent’s origin to be an important factor in landscape preferences (e.g., Brody et al., 2004; Scott, 2002). Buijs et al. (2006), identified differences between native Dutch respondents and immigrants from Mediterranean Islamic countries. Immigrants expressed particularly low levels of preference for wild and unmanaged landscapes. They also theorized that a familiarity with Dutch landscapes might influence the preferences of all immigrant groups, not only those from the specific region. This notion could also be applicable in our study, given that the Portuguese differed significantly in their preferences when compared to a mixed group of foreigners without considering their country of origin. In the future, it would be interesting to relate the landscape preferences of first-generation newcomers to Alentejo with those of a second generation born in Portugal to assess the “acculturation” of landscape preferences. Most landscape users with university degrees who look at the Alentejo as a place to live, such as inhabitants, new rural habitants, and second-home residents who spent their childhoods in Portugal, prefer identity landscapes. Education level considerably influences preferences for identity landscapes. Users with university degrees are more attracted to identity landscapes. These results may mean that awareness of regional identity as expressed in landscapes increases in line with the formal education of its users. Looking from another angle, the public’s lack of general knowledge of the identity qualities innate to certain landscapes, particularly when the economy is under pressure, might potentially lead to lower levels of preference for these types of landscape, which might thereby endanger any protective initiatives related to these landscapes. What is certain is that education’s impact on preferences for landscape invested with high cultural values is a subject that should be explored in more depth. Studies relating education and landscape preferences have tended to focus more on the influence of environmental education on landscape preferences. In the study of Kaltenborn et al. (2009) in Norway, the eco-centric values of respondents were correlated with preferences for cultural landscapes and higher education. Some previous studies have identified education as having a positive influence on preferences for unman-

362

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

Table 5 Preference predictors for landscape clusters resulting from MLR. The parameter estimates table shows the logistic coefficient (B) for each predictor variable for each alternative category of the landscape clusters. The open natural landscape and hunting were used as the reference categories. Parameter estimates Landscape clustersa

B

Std. error

Wald

df

Sig.

Exp (B)

95% Confidence interval for exp(B)

1 1 0 1 1 1 1 1 1 1 0 1 0

.000 .833

.948

.576

1.561

.432 .565 .654 .000 .085 .631 .443

1.548 .770 1.225 .029 3.999 .796 .680

.521 .315 .504 .012 .827 .313 .253

4.602 1.879 2.980 .069 19.335 2.022 1.822

.611

.866

.498

1.507

1 1 0 1 1 1 1 1 1 1 0 1 0

.246 .026

1.901

1.082

3.341

.007 .781 .003 .000 .017 .109 .614

5.559 1.180 5.091 .088 8.337 2.544 1.381

1.581 .367 1.730 .029 1.465 .813 .394

19.541 3.796 14.984 .266 47.434 7.967 4.836

.429

1.281

.694

2.364

1 1 0 1 1 1 1 1 1 1 0 1 0

.063 .001

2.519

1.473

4.307

.984 .657 .036 .000 .343 .393 .618

1.011 .813 .358 .010 2.192 .656 1.284

.330 .325 .137 .004 .433 .249 .481

3.099 2.030 .935 .028 11.090 1.727 3.429

.188

1.468

.829

2.599

Bound Identity Education User group

Gender Production Education User group

Gender Aesthetic Education User group

Gender a

Intercept Non-university degree University degree Inhabitants New rural inhabitants Land managers Hunters Second-home residents Regular visitors Tourists Eco-tourists Women Men

1.696 −.053 0 .437 −.262 .203 −3.531 1.386 −.228 −.386 0 −.144 0

.393 .254

18.649 .044

.556 .456 .454 .434 .804 .476 .503

.618 .330 .201 66.089 2.971 .231 .589

.282

.258

Intercept Non-university degree University degree Inhabitants New rural inhabitants Land managers Hunters Second-home residents Regular visitors Tourists Eco-tourists Women Men

−.593 .642 0 1.715 .166 1.627 −2.426 2.121 .934 .323 0 .247 0

.511 .288

1.345 4.989

.641 .596 .551 .562 .887 .582 .639

7.153 .077 8.730 18.632 5.716 2.571 .255

.313

.625

Intercept Non-university degree University degree Inhabitants New rural inhabitants Land managers Hunters Second-home residents Regular visitors Tourists Eco-tourists Women Men

.761 .924

.410 .274

3.448 11.398

.011 −.207 −1.028 −4.610 .785 −.422 .250 0 .384 0

.571 .467 .490 .528 .827 .494 .501

.000 .197 4.394 76.194 .900 .729 .249

.292

1.732

0

Bound

The reference category of clusters is: Open natural landscapes.

aged, wild or natural landscapes (Zheng et al., 2011). This influence is not clearly borne out by our study, as belonging to a specific user group, for example, being a hunter or not, is also a highly significant factor in determining a preference for unmanaged, natural landscapes. This study also points to the important differences between landscape user preferences for non-intensive olive groves compared to intensive olive groves. While non-intensive olive groves are highly appreciated on account of their identity qualities, intensive olive groves receive relatively low preference ratings, particularly by landscape consumers, and are perceived as mainly a production land cover type. These results indicate that the intensity of land management matters not only in environmental aspects (Beaufoy, 2001) but also for cultural ones and how different user groups appreciate them. Policymakers may be able to influence, in particular, the management of land cover qualities identified in the study. While most inhabitants, new rural habitants, second-home residents, tourists and eco-tourists appreciate identity assets on land covers, hunters are more interested in ecological qualities providing good hunting conditions. Identity land cover qualities are also appreciated by land managers, even if their main focus is on production. New rural

inhabitants, tourist and eco-tourists also appreciate the aesthetic and wilderness land cover qualities. The results from Alentejo differ from those of the meta-analysis of preferences across the whole of Europe (Van Zanten et al., 2014). While overall European preferences are for agricultural mosaic, in the Alentejo, this land cover type comes second to montado. This lends credence to previous studies regarding the assumption that landscape preferences are dependent on regional context (Tveit, 2009), and to address landscape specificities, regional studies are extremely important (Waltert et al., 2011). The upholding of Europe’s valuable diversity calls for contextualized solutions for land-use planning to bring out the full potential of regions (Henle et al., 2008). In general, montado is the most popular land cover type among landscape users in the Alentejo. It is a land-use system with mixed silvo-pastoral components and unique to the region. Apart from its leading position in terms of regional identity, previous research has also revealed its importance in other respects. It is recognized for its multifunctional use (Surová and Pinto-Correia, 2009), for supporting a variety of ecosystems (Bugalho et al., 2011) and is regarded as high-nature-value farmland (Pinto-Correia and Godinho, 2013). This demonstrates that montado is a land-use system in which societal values and ecological benefits are simultaneously aligned.

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

363

Table 6 Landscape cluster probability membership in percentage per predictor level. Landscape clusters Open natural landscape n = 228

Identity landscape n = 396

Production landscape n = 190

Aesthetic landscape n = 252

User groups (p = 0.000) Inhabitants New rural inhabitants Land managers Hunters Second-home residents Regular Visitors Tourists Eco-tourists

5.3 11.3 7.3 77.5 2.7 10.3 9.6 8.8

40.4 43.6 46.1 11.8 54.1 41.9 32.7 43.9

28.1 10.5 34.5 6.9 18.9 21.4 10.6 7.9

26.3 34.6 12.1 3.9 24.3 26.5 47.1 39.5

Education Level (p = 0.000) University degree Non-university degree

13.6 26.0

49.5 29.9

14.4 19.9

22.5 24.3

Gender (p = 0.000) Women Men

8.5 27.2

37.3 37.1

18.5 17.5

35.8 18.2

Age (p = 0.002) Before 1970 After 1970

23.4 17.5

35.4 40.6

19.8 13.9

21.4 28.1

Childhood residence place (p = 0.000) Alentejo Portugal Foreign country

27.0 15.3 12.9

35.6 42.4 33.9

21.3 12.8 14.5

16.0 29.5 38.7

Relation to agriculture (p = 0.000) Yes No

23.7 17.7

38.1 35.6

20.4 13.8

17.8 32.9

Maintenance and planning for this type of alignment is recognized as an important goal of multifunctional landscape planning, management and design (Gobster et al., 2007). Based on the findings of this study, landscape users in Alentejo are more attracted to agricultural systems than to forests or to uncultivated agricultural land. This may be explained by the identity qualities attached to agriculture in the region. It is as such that, unlike in previous studies (Arriaza et al., 2004), where one of the strongest predictors of landscape preference was a degree of wilderness, in the case of multi-user perspective in the Alentejo region, cultivated land is more popular. In Alentejo, cereal crop has gained a sense of societal identity since 1929 when the “Wheat Campaign” was launched, even on poor soils, thereby influencing the regional landscape and the way of life of its inhabitants (Baptista, 1993). The present survey indicates that the people in Alentejo still associate cereal crops with their regional identity, even now that the campaign is long gone. Landscape users do not present a consensus of preference for wilderness, this being more widely appreciated by the nonPortuguese users, particularly mixed forest. Even in terms of preferred aesthetic attributes, tastes in the Alentejo are markedly more taken with heterogeneous agricultural mosaics and planted pine forests than with more wild landscapes such as semi-natural areas or forested lands. Most of the forest areas in the region are planted and have no long history. Forest plantations on previously farmed land was subsidized by the European Union (reg./CEE/2080/92), and consequently, the large-scale land managers who had previously leased their land to local farmers decided to manage their forested land directly. At the same time, these areas became part of tourist hunting grounds with restricted access for other landscape users. These changes often led to the exclusion of local farmers and other local landscape users from the forested areas. Subsequently, they began to be viewed as “non-local”, merely serving a specific group of people (Carolino, 2010). This might explain why the densely forested areas of the Alentejo region are

less popular among Portuguese landscape users who are familiar with the “past” than among non-native users. Moreover, in our study, preferences for wilderness did not correlate with levels of education, as had been the case in previous studies (e.g., Van den Berg et al., 1998), but rather with origin. Having spent one’s childhood residing in Portugal presented a negative correlation with demand for wilderness, mixed forest and high shrubs in forest land when compared with having spent one’s childhood in a foreign country. This study conducted a quantitative photo-questionnaire with qualitative components at a personal and individual level. This type of methodology has both advantages and disadvantages. To conduct a quantitative evaluation of region-specific landscape preferences is time-consuming and also requires considerable financial resources for the transportation of researchers during the photopreparation process as well as to conduct the survey. However, the primary advantage of the qualitative approach is that it is ideally suited to the collection of rich qualitative data, including quotations from individuals, as well as the presentation to respondents of questions with a high cognitive burden. Moreover, respondents were able to freely explain their preferences to shed further light on the landscape attributes they appreciated most. The combination of quantitative and qualitative data provided deeper insight into understanding land users’ preferences for Mediterranean landscapes than the purely quantitative data would have. In areas where landscape preference issues have not been explored at length, this more personal survey with its qualitative dimension would appear to be an important step in gaining experiential knowledge about landscape appreciation, which could subsequently be followed in the future by other, less laborintensive assessments and validations. One such option might be an online survey (Brown and Mortimer, 2014). The photographs used in the survey were of single LCL class, which results in a considerable simplification of real landscape sit-

364

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365

uations. However, in comparison with real landscape scenarios, the LCL class evaluation allows for a more direct comparison of the same evaluation in the future as well as considering trends in user appreciation level with regard to potential land cover changes. In addition, recent research shows that the link between land cover and preference data may be applied to the formulation of spatial representations of landscape demands (Carvalho-Ribeiro et al., 2013). One of the potential limitations of the presented landscape typology might be in its translation into real landscape situations. Landscape typology based purely on user preferences rather than on expert or statistical judgement might result in greater difficulties in identifying real locations with the desired landscape. In this study, preferences were expressed as a percentage of the respondents’ choice instead of using scale evaluation of each photo, in the interests of making the questionnaire less challenging for respondents and more user-friendly. Consequently, the data were nominal and would benefit from more complex statistical analysis with an initial quantification process. Our methodology might be improved by using the Likert scale, which generates easilyquantifiable data.

6. Conclusions The analysis presented in this paper shows that the way that people engage with landscape and the functional expectations they have of it significantly influence their land cover preferences. One may identify differences not only between land managers as landscape production users and others as landscape consumers, but land cover qualities and land cover preferences also vary between those who are largely consumers. An informed understanding of this diversity in landscape preferences between user groups is needed to articulate effective landscape management strategies capable of ensuring the landscape’s multi-functionality—both in policy formulation and in planning. Furthermore, at the farm management level, it makes it possible for the decision-maker to assess what impacts the present and planned management will have on users’ satisfaction with the resulting landscape. For the purposes of hunting or recreational activities, this may be highly relevant at the single farm or local landscape level. Moreover, this knowledge also opens up the possibility of informing users, in accordance with their landscape expectations, about where and to what extent their expectations may be met by a given landscape. The knowledge obtained by the analysis that has been undertaken is thus a step forward in facilitating a more accurate matching-up of rural landscape demands with supply. Furthermore, the results of this study bear out the importance of regional-level assessment of landscape preferences, due to its ability to detect sensitive regional nuances. The identification of the landscape attributes that are specifically valued in a particular location may prove to be of assistance in maintaining the highly prized heterogeneity of Europe’s landscape. The question may be raised as to what extent the complexities in landscape preferences that this study has identified are directly related to and caused by the underlying complexity of Mediterranean landscapes. In more complex landscapes, which support multiple uses in the same area, the diversity of user expectations may be expected to be higher than in more homogeneous landscapes. The comparison of the results presented in this paper with those of preference studies in other regions of Europe seems to indicate that this would be the case. As a step toward usefulness of science for policy, this study shows that if the surveys’ landscape variations are directly linked to changes in agricultural or forest land cover, then a more direct

relationship to the way that policy impacts on landscape may be established (Styers et al., 2010). Finally, the assessment of landscape user preferences should be seen as just one step, albeit an important one, on an interactive and circular path linking scientific knowledge, planning, planning implementation and assessment of the consequences of said planning. Any specific local land-use management will need to involve the participation of local citizens and achieve acceptable trade-offs. Acknowledgments The authors would like to thank the INALENTEJO programme and the regional administration bodies, namely the Regional Development Commission (CCDR Al) in partnership with the regional ministry of agriculture (DRAP Al), which all together sponsored the research project ROSA for surveying the land cover pattern preferences by different landscape users in the Alentejo region in southern Portugal. Acknowledgements are also due to the Portuguese Science Foundation (FCT) for the funding provided for the post doc scholarship of the leading author with the reference SFRH/BPD/77649/2011. This work was funded by FEDER Funds through the Operational Programme for Competitiveness Factors—COMPETE and National Funds through FCT—Foundation for Science and Technology under the Strategic Project PEstC/AGR/UI0115/2011. Moreover, acknowledgements are due to the PEGASUS project receiving funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 633814. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.landusepol.2016. 02.026. References ˜ ˜ ˜ Ruiz-Aviles, P., 2004. Assessing the Arriaza, M., Canas-Ortega, J.F., Canas-Madue no visual quality of rural landscapes. Landscape and Urban Plann. 69, 115–125. Baptista, F.O., 1993. A Política Agrária do Estado Novo. Porto, Afrontamento. Barroso, F., Pinto-Correia, T., Ramos, I.L., Surová, D., Menezes, H., 2012. Dealing with landscape fuzziness in user preference studies: Photo-based questionnaires in the Mediterranean context. Landscape Urban Plann. 104 (3–4), 329–342, http://dx.doi.org/10.1016/j.landurbplan.2011.11.005. Beaufoy, G., 2001. The environmental impact of olive oil production in the European Union: practical options for improving the environmental impact. Eur. Comm. [online] Available from: http://ec.europa.eu/environment/ agriculture/pdf/oliveoil.pdf (accessed 19.10.15.). Berninger, K., Adamowicz, W., Kneeshaw, D., Messier, C., 2010. Sustainable forest management preferences of interest groups in three regions with different levels of industrial forestry: an exploratory attribute-based choice experiment. Environ. Manage. 46 (1), 117–133. Brody, S.D., Highfield, W., Alston, L., 2004. Does location matter? Measuring environmental perceptions of creeks in two San Antonio watersheds. Environ. Behav. 36 (2), 339–350. Brown, P., Mortimer, C., 2014. Econometric analysis of landscape preferences in Canterbury, New Zealand. Econ. Res. Int., 1–12, http://dx.doi.org/10.1155/ 2014/259471. Bugalho, M.N., Caldeira, M.C., Pereira, J.S., Aronson, J., Pausas, J.G., 2011. Mediterranean cork oak savannas require human use to sustain biodiversity and ecosystem services. Front. Ecol. Environ. 9, 278–286, http://dx.doi.org/10. 1890/100084. Buijs, A.E., Elands, B.H.M., Langers, F., 2006. No wilderness for immigrants: cultural differences in images of nature and landscape preferences. Landscape Urban Plann. 91, 113–123. Burton, R.J.F., Kuczera, C., Schwarz, G., 2008. Exploring farmers cultural resistance to voluntary agri-environmental schemes. Sociol. Rural. 48, 16–37. Carolino, J., 2010. The social productivity of farming: a case study on landscape as a symbolic resource for place-making in Southern Alentejo, Portugal. Landscape Res. 35 (6), 655–670, http://dx.doi.org/10.1080/01426397.2010.519437. Carvalho-Ribeiro, S., Migliozzi, A., Incerti, G., Pinto Correia, T., 2013. Placing land cover pattern preferences on the map: bridging methodological approaches of landscape preference surveys and spatial pattern analysis. Landscape Urban Plann. 114, 53–68, http://dx.doi.org/10.1016/j.landurbplan.2013.02.011.

D. Surová, T. Pinto-Correia / Land Use Policy 54 (2016) 355–365 Carvalho-Ribeiro, S.M., Pinto-Correia, T., Paracchini, M.L., 2016. Addressing the social landscape dimensions: The need for reconciling cross scale assessments for capturing Cultural Ecosystem Services (CES). Land Use Policy, http://dx.doi. org/10.1016/j.landusepol.2015.12.022, Available online 6 January 2016, ISSN 0264-8377. CM, Recommendation CM/Rec, 2008, 3 of the Committee of Ministers to member states on the guidelines for the implementation of the European Landscape Convention Bruxelles. Council of Europe, 2000. European landscape convention, Florence, CETS No. 176. Counc. Eur. Strasbg. Davis, H.Ch., Michelle, C., 2011. Q methodology in audience research: bridging the qualitative/quantitative divide? Participations 8 (2), 559–593. Dax, T., 2014. Shaping rural development research in Europe: acknowledging the interrelationships between agriculture, regional and ecological development. Stud. Agric. Econ. 116, 59–66. European Commission, 2010. The CAP towards 2020: meeting the food, natural resources and territorial challenges of the future. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the regions. COM (2010) 672 final Brussels. Giannoccaro, G., Berbel, J., 2013. Farmers stated preference analysis towards resources use under alternative policy scenarios. Land Use Policy 31, 145–155. Gobster, P.H., Nassauer, J.I., Daniel, T.C., Fry, G., 2007. The shared landscape: Whatdoes aesthetics have to do with ecology? Landscape Ecol. 22 (7), 959–972, http://dx.doi.org/10.1007/s10980-007-9110-x. Greenacre, M., 2006. Multiple Correspondence Analysis and Related Methods. Taylor and Francis, Abingdon, 607 p. Henle, K., Alard, D., Clitherow, J., Cobb, P., Firbank, L., Kull, T., McCracken, D., Moritz, R.F., Niemel, J., Rebane, M., Wascher, D., Watt, A., Young, J., 2008. Identifying and managing the conflicts between agriculture and biodiversity conservation in Europe—a review. Agric. Ecosyst. Environ. 124, 60–71. Hofmann, M., Westermann, J.R., Kowarik, I., Van der Meer, E., 2012. Perceptions of parks and urban derelict land by landscape planners and residents. Urban For. Urban Greening 11 (3), 303–312. Holmes, J., 2012. Cape York Peninsula, Australia: a frontier region undergoing a multifunctional transition with indigenous engagement. J. Rural Stud. 28, 1–14. Howley, P., 2011. Landscape aesthetics: assessing the general publics’ preferences towards rural landscapes. Ecol. Econ. 72, 161–169, http://dx.doi.org/10.1016/j. ecolecon.2011.09.026. INE, 2014. Portugal em números 2013. In: Portugal in Figs. 2013. Instituto Nacional de Estatística, I.P., Lisboa—Portugal, 58pp. Johansson, S., Loetscher, M., Gennaio, M.-P., Dreger, F., Brouwer, F., Dax, T., Meredith, D., Dernfalk, J., Ohlson, B., 2012. Strategic research Agenda Deliverable 3.2 of the ERA-NET RURAGRI. Jones, P.J., Andersen, E., Capitani, C., Carvalho-Ribeiro, S., Griffiths, G., Loupa-Ramos, I., Madeira, L., Mortimer, S., Paracchini, M.L., Pinto-Correia, T., Schmidt, A., Simoncini, R., Wascher, D., 2015. The EU societal awareness of landscape indicator: a review of its meaning, utility and performance across different scales. Land Use Policy, http://dx.doi.org/10.1016/j.landusepol.2015.01.038. Jones, M., Stenseke, M., 2011. The European landscape convention challenges of participation. In: Landscape Series. Springer, Netherlands, 326 pp. Junge, X., Lindemann-Matthies, P., Hunziker, M., Schüpbach, B., 2011. Aesthetic preferences of non-farmers and farmers for different land-use types and pro-portions of ecological compensation areas in the Swiss lowlands. Biol. Conserv. 144 (5), 1430–1440, http://dx.doi.org/10.1016/j.biocon.2011.01.012. Kaltenborn, B., Andersen, O., Nellemann, C., 2009. Amenity development in the Norwegian mountains Effects of second home owner environmental attitudes on preferences for alternative development options. Landscape Urban Plann. 91, 195–201. Marsden, T., Sonnino, R., 2008. Rural Development and the regional state: denying multifunctional agriculture in the UK. J. Rural Stud. 24, 422–431. Natori, Y., Chenoweth, R., 2008. Differences in rural landscape perceptions and preferences between farmers and naturalists. J. Environ. Psychol. 28, 250–267. Paracchini, M.L., Pinto-Correia, T., Loupa-Ramos, I., Capinani, C., Madeira, L., 2015. Progress in indicators to assess agricultural landscape valuation: how and what is measured at different levels of governance. Land Use Policy. http://dx. doi.org/10.1016/j.landusepol.2015.05.025. Paracchini, M.L., Capitani, C., Schmidt, A.M., Andersen, E., Wascher, D.M., Jones, P.J., Simoncini, R., Carvalho Ribeiro, S., Griffiths, G.H., Mortimer, S.R., Madeira, L., Ramos, I., Pinto Correia, T., 2012. Measuring societal awareness of the rural agrarian landscape: indicators and scale issues. In: Report from the Framework Contract Project FRAGARIA. Joint Research Centre, Italy, 150 pages.

365

Patton, M.Q., 2002. Qualitative Research and Evaluation Methods, 3rd ed. Sage, Thousand Oaks, CA. Pinto-Correia, T., Barroso, F., Surová, D., Menezes, H., 2011. The fuzziness of Montado landscapes: progress in assessing user preferences through photo-based surveys. Agrofor. Syst. 82 (2), 209–224. Pinto-Correia, T., Carvalho-Ribeiro, S., 2012. The Index of Function Suitability (IFS): a new tool for assessing the capacity of landscapes to provide amenity functions. Land Use Policy 29 (1), 23–34. Pinto-Correia, T., Godinho, S., 2013. Chapter 4 changing agriculture-changing landscapes: What is going on in the high valued montado landscapes of southern Portugal? In: Ortiz-Miranda, D., Moragues-Faus, A., Arnalte-Alegre, E. (Eds.), Agriculture in Mediterranean Europe: Between Old and New Paradigms (Research in Rural Sociology and Development), vol 19. Emerald Group Publishing Limited, Bingley, pp. 75–90. Pinto-Correia, T., Kristensen, L., 2013. Linking research to practice: the landscape as the basis for integrating social and ecological perspectives of the rural. Landscape Urban Plann. 120, 248–256. Pinto-Correia, T., Menezes, H., Barroso, F., 2014. The landscape as an asset in Southern European fragile agricultural systems: contrasts and contradictions in land managers attitudes and practices. Landscape Res. 39 (2), 205–217, http://dx.doi.org/10.1080/01426397.2013.790948. Pohl, C., 2008. From science to policy through transdisciplinary research. Environ. Sci. Policy 11 (8), 46–53. Primdahl, J., Andersen, E., Swaffield, S., Kristensen, L., 2013. Intersecting dynamics of agricultural structural change and urbanisation within European rural landscapes: change patterns and policy implications. Landscape Res. 38 (6), 799–817, http://dx.doi.org/10.1080/01426397.2013.772959. Primdahl, J., Swaffield, S., 2010. Globalisation and Agricultural Landscapes: Change Patterns and Policy Trends in Developed Countries. Cambridge University Press, Cambridge. Purcell, M., 2006. Urban democracy and the local trap. Urban Stud. 43 (11), 1921–1941. Rambonilaza, M., Dachary-Bernard, J., 2007. Land-use planning and public preferences: what can we learn from choice experiment method? Landscape Urban Plann. 83, 318–326. Rogge, E., Nevens, F., Gullink, H., 2007. Perception of rural landscapes in Flandres: looking beyond aesthetics. Landscape Urban Plann. 82 (4), 159–173. Scott, A., 2002. Assessing public perception of landscape: the LANDMAP experience. Landscape Res. 27, 271–295. Selman, P., 2012. Sustainable Landscape Planning. In: The Reconnection Agenda. Earthscan, Routledge, London and NY, 162 pp. Styers, D.M., Chappelka, A.H., Marzen, L.J., Somers, G.L., 2010. Developing a land-cover classification to select indicators of forest ecosystem health in a rapidly urbanizing landscape. Landscape Urban Plann. 94 (3-4), 158–165. Surová, D., Pinto-Correia, T., 2008. Landscape preferences in the cork oak Montado region of Alentejo, southern Portugal: searching for valuable landscape characteristics for different user group. Landscape Res. 33 (3), 311–330. Surová, D., Pinto-Correia, T., 2009. Use and assessment of the ‘new’ rural functions by land users and landowners of the Montado in southern Portugal. Outlook Agric. 38 (2), 189–194. Swanwick, C., 2009. Society’s attitudes to land and landscape. Land Use Policy 26S, S62–S75. Tudor, C.A., Ioj˘a, I.C., Pˇatru-Stupariu, I., Nita, M.R., Hersperger, A.M., 2014. How successful is the resolution of land-use conflicts? A comparison of cases from Switzerland and Romania. Appl. Geogr. 47, 125–136. Tveit, M.S., 2009. Indicators of visual scale as predictors of landscape preference; a comparison between groups: environmental and landscape change: addressing an interdisciplinary agenda. J. Environ. Manage. 90 (9), 2882–2888. Van den Berg, A.E., Vlek, C.A.J., Coeterier, J.F., 1998. Group differences in the aesthetic evaluation of nature development plans: a multilevel approach. J. Environ. Psychol. 18, 141–157. Van Zanten, B.T., Verburg, P.H., Koetse, M.J., Van Beukering, P.J.H., 2014. Preferences for European agrarian landscapes: a meta-analysis of case studies. Landscape Urban Plann. 132, 89–101, http://dx.doi.org/10.1016/j.landurbplan. 2014.08.012. Waltert, F., Schulz, T., Schläpfer, F., 2011. The role of landscape amenities in regional development: evidence from Swiss municipality data. Land Use Policy 28 (4), 748–761. Zheng, B., Zhang, Y., Chen, J., 2011. Preferences to home landscape: wildness vs neatness. Landscape Urban Plann. 99, 1–8.