Change Adaptation Socioecol. Syst. 2017; 3: 93–102

1 IntroductionFollowing a legal mandate that had been in force since 2008, the Regional Administration of Galicia (Spain) set, in the beginning of 2015, the requirements for a map of landscape units of the region that would constitute the basis of a larger project to create a Landscape Catalogue. Galicia is a region in NW Spain with a territory approximately the size of the Netherlands (29 574 km²), highly fragmented land use and land property systems and heterogeneous topography. The terms of reference required a rather detailed scale in the final product of 1:25,000 and the requirement to finish the work in less than three months. The urgency was justified, in part, because a new land-use act that centralized some of the previously municipal competencies related to land-use zoning was expected to be passed soon: the regional ministry was very interested, therefore, not only in having a map of landscape units, but particularly of protected landscape areas before the land-use act was in force.

An advisory expert panel, formed by sixteen experts coming both from the academic and public sector, was formed in order to provide technical guidance to the project. The selection of experts was based on the number of recent publications in the broad field of landscape studies. In addition, the expert panel pretended to be multidisciplinary and balanced by combining experts that understand landscape mainly as a form and culture (architects, geographers, historians, etc.) and experts that pay more attention to the functional aspects of landscape (biologists, engineers, etc.). The role of the expert panel was to contribute to the following four tasks:1. Reviewing and suggesting types of landscape to be

mapped.2. Writing a document summarizing the main landscape

values and their conservation degree in the region.3. Writing a document summarizing the main drivers

and dynamics that influence the different types of landscapes.

4. Reviewing the final document.

https://doi.org/10.1515/cass-2017-0008received November 25, 2016; accepted September 5, 2017

Abstract: In the beginning of 2015, the Regional Administration of Galicia (NW Spain) set the requirements for a map of landscape units: it had to be produced in less than 3 months, it should cover the whole territory of the region (29,574 km²), and it should be useful for management at a scale of 1:25,000. With these objectives in mind, we pro- posed a semi-automatic mapping methodology entirely based on the use of free software (GRASS GIS) and already available cartographic information. Semi-automatic classification of different land-use patterns was at the heart of the proposed process. Consultation with experts of different academic background took place along the project. This consultation process allowed to identify both problems and opportunities. As it could be expected, the diverse epistemic community represented by the expert panel implied that one of the main challenges was to reach consensus on the understanding of the concept of landscape and the decisions leading to the mapping methodology proposed in this paper. This initiated a very interesting debate that, in our view, was centred around three main issues: the approach to the landscape, the purpose of the mapping exercise, and the ability to include subjectivity into the analysis.

Keywords: landscape units, landscape mapping, geographic information systems, semi-automatic classification, epistemic community

Research Article

Eduardo Corbelle-Rico*, José María Tubío Sánchez

Mapping landscape units in Galicia (Spain): A first step for assessment and management?

*Corresponding author: Eduardo Corbelle-Rico, Land Laboratory, Department of Agricultural and Forest Engineering. Universidade de Santiago de Compostela. Benigno Ledo s/n, 27002 Lugo, Spain, E0mail: eduardo.corbelle@usc.esJosé María Tubío Sánchez, Land Laboratory, Department of Agricul-tural and Forest Engineering. Universidade de Santiago de Compo-stela. Benigno Ledo s/n, 27002 Lugo, Spain

Open Access. © 2017 Eduardo Corbelle-Rico, José María Tubío Sánchez, published by De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 License.

94   E. Corbelle-Rico, J.M. Tubío Sánchez

To achieve these goals, the members of the expert panel were divided into three working groups and each group received the assignment of paying attention to one particular area within the region. These three subdivisions of the territory of Galicia were defined according to major landscape changes in the last decades (urbanization, intensification of agriculture and land abandonment), as proposed by Corbelle Rico and Crecente Maseda [1]. One of the experts in each group was assigned the role of rapporteur and was in charge of writing, with the collaboration of the other group members, the required documents. A total of two meetings of the complete panel were held. In the first meeting, the goals and methodology of the whole project were introduced to the panel, and attention was concentrated on task 1. After that, working groups worked on tasks 2 and 3 along the course of three months. Finally, all members received a copy of the final document and a second general meeting was held to improve it with their suggestions.

Classifying and mapping landscape units is often the frame for landscape management [2]. Obviously, the basic assumption behind any landscape mapping exercise is that distinct categories (landscape types) can be identified, showing some degree of internal homogeneity concerning their biophysical, cultural and aesthetic characteristics that make them different from other categories [3]. While different sets of characteristics may be considered for classification purposes, it is also assumed that the ones finally selected should be relevant from the management point of view, and each landscape type should be accompanied by a brief description outlining the aspects that are most relevant for management purposes [e.g. 4]. The concrete representation of a landscape type on the earth surface is named, for the purposes of this work, a landscape unit. Following Kienast et al. [5], landscape units can be defined as “medium-scale excerpts of the globe’s surface, shaped by nature and humans and perceived by people”. Different units may be classified as being of the same type when they have the same character: “a distinct, recognisable and consistent pattern of elements in the landscape that makes one landscape different from another, rather than better or worse” [6]. The concept of landscape character has been adopted by a large scientific community and is recurrent in contemporary landscape mapping exercises [e.g. 7, 8]. Of course, the aforementioned “recognisable pattern of elements” may vary with the chosen scale. Scale of work is a crucial decision to be made depending on the purposes of the project that may completely change the selected approach.

In the past, the process of mapping landscape units has been usually based on expert (i.e., “subjective”) judgment. Maps were fundamentally based on subjective appreciation and interpretation of different cartographic sources by experts that should then manually delineate the units on a clean sheet of paper. While subjective methods often result in rich and detailed maps, they demand a lot of time to be prepared and can result obscure from the end-user perspective and difficult to be revised by other experts. They also present very important problems to incorporate new (e.g. updated) data [8]. In recent decades, classification procedures based on quantitative statistical methods (i.e., “objective” methods) are used in the vast majority of landscape mapping exercises, as they are usually much easier to replicate, update, and document. These methods make use of state-of-the-art information technologies like geographic information systems, remote sensing, and machine learning algorithms [8, 9].

Most works in the literature use a relatively simple approach, including a low number of variables that usually is in the range of 3–5: climate and topography, complemented with protected spaces, potential vegetation and current land use [3]; climate, soils, topography, potential vegetation and current land use [4]; elevation, slope, ecosystem classification and geology [10]; elevation, land use, soil types and satellite imagery [11]. Less frequently, some authors present a seemingly large number of variables but end up using just a few for the mapping exercise: for example, Mücher et al. [8] combine abiotic factors (climate, geomorphology, hydrology, soil type), biotic factors (vegetation and fauna), and cultural factors (land-use/land-cover and landscape structure), although they use just climate, elevation, geology and land-use/land-cover for the mapping. According to Vallés et al. [12], landscape classifications produced in Spain are usually based on topography, land use and visibility. Particularly, works carried out at regional level usually base the topographic analysis on the main topographic domains (mountain ranges, plains…) and not on more detailed landforms (valley bottoms, hillsides, peaks). Visibility analysis are not really common at this (regional) level, or are very general.

In this document we aim to outline the main methodological decisions for the creation of a landscape map of Galicia. We describe the main data sources and procedural methodologies and discuss them in the light of the initial requirements. Rather than discussing the methodological aspects in detail, the paper focus the discussion on the interaction with the assessing expert panel and the regional Administration, outlining the main debates aroused by the proposed methodological approach and their implications for future projects.

Mapping landscape units in Galicia (Spain)   95

2 Materials and methodsBy and large, time and resource constraints conditioned decisions on data sources and methods. Accordingly, we designed our mapping approach so it could be: relatively simple (i.e., not unnecessarily complicated); based on a relatively low number of —already existent— cartographic data sources; easy to reproduce by other teams; easy to update; and well documented for the public or additional experts interested in the process. The whole process was carried out using GRASS GIS 7.0 [13], using the additional module GeoPAT (Geospatial Pattern Analysis Toolbox) [14], both of which are free software applications (as defined by the Free Software Foundation: https://www.gnu.org/philosophy/free-sw.en.html). In our view, the use of free software contributes to the reproducibility [15, 16] of the methodology for two reasons: because they can be legally used by any institution, at no cost in the case of the two applications mentioned here; and, most importantly, because their computer code is always provided, thus avoiding the “black box” situation typical of some proprietary applications.

Owing to the objective of simplicity, the final map would result from the overlay of three intermediate layers: topographic domains, land-use/land-cover patterns, and climatic (temperature) domains. Four main data sources were selected as the input for the process: a digital elevation model with 25 m/pixel resolution; a digital land-use/land-cover map at 1:25,000 scale (Sistema de Información de la Ocupación del Suelo de España, SIOSE); the Natura 2000 Regional Master Plan at scale 1:25,000; and map of thermotypes of Galicia at undefined scale. The first and second sources are produced and distributed by the Spanish Geographic Institute. The Natura 2000 Master Plan is produced by the regional Administration [17], and the thermotype map was published by Rodríguez Guitián and Ramil Rego [18]. The latter was used in its original form and not subjected to any additional process. Land-use/land-cover classes in SIOSE and Natura 2000 Master Plan were combined to produce a map with a reduced number of land-use/land-cover categories: agricultural areas, vineyards, plantation forests, deciduous broadleaved forests, dry shrublands, urban areas, mines and quarries, water bodies, wetlands.

2.1 Main topographic domains

The classification of main topographic domains within the region followed that proposed by Ramil Rego et al. [19]. This divides Galicia in four main domains: Cantabric-

Atlantic littoral, sublittoral valleys, mountain ranges, and interior plains and valleys. We added a fifth category to classify river canyons in a different class. A completely new version of the map was generated starting with the automatic segmentation of the terrain elevation and terrain slope layers, combined. This resulted in more than 12 000 segments of homogeneous altitude and slope (Fig. 1), which were manually classified into the aforementioned five classes . Generally, areas located under 100 m of elevation were considered to be included in the “littoral” class, while areas located above 600–700 m (400–500 m if they were close to the sea) and with higher slopes were considered to be “mountain areas”.

2.2 Land-use/land-cover domains

Classification of land-use/land-cover patterns implies a step further the classification of land-use/land-cover. In this approach, each patch of homogeneous land use is treated as a landscape element. The combination of such elements, when considered from a patch/matrix/corridor perspective, help to define a pattern related to the kind of environmental services provided by each tract of land [20].

Considering that Galician landscapes are deeply rooted in agricultural and forest management practices, several of the classes were defined as different kinds of agro-ecosystems. In accordance with the very different nature of the ecosystem services that they provide, these areas were further divided into several sub-categories: areas of intensive agricultural production, dominated by arable land and intensive forage production; areas of extensive agricultural production, in which small parcels of cultures or pastures coexist with extensively grazed shrublands and small stands of deciduous hardwood forest; and areas where intensive agriculture, cattle husbandry and plantation forests are closely intertwined. The list of land-use/land-cover patterns considered includes several other categories (Table 1).

One or more training areas were selected for each class of land-use/land-cover pattern, consisting of a circle of 1 km in diameter (Fig. 2), to be used in the automatic search and classification process. For ideally “pure” classes (i.e. those dominated by just one class of land-use/land-cover) the number of training samples selected was just one. An statistical measure of class co-occurrence was calculated for each training sample. Co-occurrence measures how many times a given class appears next to each other within the circular window [14] and is, therefore, very different from a simple class-frequency count.

96   E. Corbelle-Rico, J.M. Tubío Sánchez

A map of similarity with each of the land-use/land-cover patterns, using a moving window of 1 km of diameter, was produced using Shannon distance [21], which Jasiewicz et al. [9] consider as closer to human visual perception of similarity. When several training areas were used for the same land-use/land-cover pattern class, similarity maps were produced for each one and

then averaged to produce a single similarity map. Finally, each pixel in the map was assigned to the land-use/land-cover pattern class with a higher similarity value. The only exception to this process were the historical sites and water bodies, which were directly translated from the original maps to the land-use/land-cover pattern map (Fig. 3).

Figure 1. An example of the automatic segmentation that constituted the base for the classification of main topographic domains. The segment limits (red line) show the effect of the original resolution of the DEM. The areas in blue are an example of areas classified as “river canyons” in the area of the river Eume. The figure uses a topographic map obtained via WMS as reference.

Table 1. Land-use/land-cover pattern classes.

Land-use/land-cover pattern1 Dry shrubland2 Peatland3 Deciduous forests4 Intensive agro-ecosystems4.1 Areas of intensive agriculture4.2 Mosaic of forest plantations and intensive agriculture4.3 Forest plantations4.4 Vineyards5 Extensive agro-ecosystems6 Residential areas6.1 Urban areas6.2 Low density residential areas6.3 Historical sites *7 Open pit mining areas8 Water bodies *

* Areas not generated by pattern classification

Mapping landscape units in Galicia (Spain)   97

3 ResultsThe overlay of the three main layers (topographic domains, land-use/land-cover patterns, and thermotypes, Figs. 4-6) resulted in a map with around 28 000 different landscape units with an average area of about 1 km². Each landscape unit was, then, classified according to the categories of each map. The system theoretically allows end-users to

select just one of these layers (or a combination of two) if this suits their needs better than the final map with the three categories combined. The number of actual combinations (landscape types) was 258 (although the theoretical number of possible combinations was larger). The map is available for the public in a web-gis system provided by the regional Administration (available at: http://mapas.xunta.es/visores/paisaxe).

Figure 2: An example of training area for the “Mosaic of forest plantations and intensive agriculture” land-use/land-cover pattern.

Figure 3: An example of a Historical site directly translated into the land-use/land-cover pattern map.

98   E. Corbelle-Rico, J.M. Tubío Sánchez

Figure 4: Map of main topographic domains (following Ramil Rego et al. (2005).

Figure 5: Map of land-use/land-cover patterns.

Mapping landscape units in Galicia (Spain)   99

4 DiscussionOverall, we believe that the mapping exercise briefly described in this paper did comply with most of the initial requirements set up in the project’s terms of reference. Obviously, limitations in the time available led to a very straight-forward process of decision making, and this means that several methodological decisions probably left much space for improvement.

In spite of focusing on the concrete aspects of the proposed classification methodology, it is probably much more interesting to discuss its interaction with the two main agents also involved: the expert panel designated by the Regional Administration, and the Regional Administration itself. The expert panel had as its main function to contribute to improve the decision-making process and the final document (the Landscape Catalogue of which the landscape map formed part), in particular, by identifying landscape values and drivers of landscape change. The Regional Administration, on the other hand, established the requirements and time constraints. What is interesting to discuss here is how these two main actors perceived the methodology here proposed to classify landscape units. We believe this will provide lessons on the weaknesses and opportunities that should be taken into

account, from the very beginning, when designing and implementing this kind of projects. We found out that the main challenge is, on the one hand, reaching consensus on the understanding of the concept of landscape when explaining the proposed methodology to a diverse epistemic community represented by the expert panel. On the other hand, the Regional Government imposed some requirements on scale and on the participatory process, that probably undermined the scope and ability of the landscape units to be the base of a landscape management policy.

4.1 The remarks of the expert panel

The expert panel was composed by sixteen members with very different academic backgrounds and with practice experience in different private or public sectors. Their main task was to provide feedback to the mapping exercise (e.g., by proposing potential landscape types to be identified) as well as to add content to the whole project (e.g. by identifying landscape values and main drivers of landscape change). However, the panel consistently focused their remarks on three main aspects: the selected approach for the mapping exercise; the purpose of the

Figure 6: Map of main climatic domains (thermotypes), as defined by Rodríguez Guitián & Ramil Rego (2007).

100   E. Corbelle-Rico, J.M. Tubío Sánchez

resulting map; and the role of subjectivity in landscape in the context of the proposed methodology.

On the approach. The methodology described in this paper is largely based on landscape ecology concepts and fundamentally equates the concepts of “landscape” and “territory”. This was found troubling by some panel members, who suggested that a more perceptive approach to landscape should be used, instead. In particular, a lot of attention was paid to specific landscape elements (e.g. historical buildings) as, it was argued, these were missing from the moment landscape units were defined from a territorial perspective.

On the purpose. In its present form, the proposed methodology is largely based on the assumption that landscape units should be fundamentally defined according to functional aspects —i.e., the type of ecosystem services provided. Consequently, the approach emphasizes the need of understanding the kind of processes (e.g. land use changes) that take place in the landscape. Landscape policy, therefore, should focus on management of landscape processes rather than on zoning regulations and protection of particular elements by the use of buffer zones. However, this confronted the view of several panel members, who understood that the main goal of landscape policy should be regulatory, in order to protect “special” landscapes.

On the subjectivity. Some members of the panel suggested that the proposed methodology was unable to capture subjective aspects of landscape (e.g. colour combinations, odours…). This is very closely linked to the first remark, and is related to the fact that historic or natural elements that are present in the landscape usually have very powerful social values attached that cannot be easily captured with an approach like the one proposed.

Considering these main remarks, it become apparent that, by focusing the discussion on information sources or the ability of the methodology to capture certain types of landscape, we lost a great opportunity to frame the debate in a broader perspective. In fact, although the debate was apparently at the technical level, the specific details of the methodology were never discussed. What was being indirectly contested under the three remarks was, in essence, the concept of landscape underlying the methodology. The concept of landscape is still a contested one, albeit often not explicitly, as it is used by different disciplines that traditionally do not maintain a fluent dialogue among them. In practical terms, two main contending groups of disciplines could be identified: engineering, biology and earth sciences usually understand landscape from a functional perspective; disciplines like history, architecture, or in

general disciplines in the realm of the humanities, usually approach the study of landscape focusing on forms and the symbolic values attached to it [22], and therefore understand landscape as a subjective phenomena [23].

While there is an increasing acknowledgement about the need of an integrated approach to study landscape [24], the cleavage between sciences and humanities has not yet been overcome. Barriers to integration like the difficulty of coping with different academic traditions and the lack of common terminology cannot be overcome if a considerable portion of time is not dedicated to create a consensus on the terminology [25]. In the case presented here, it is evident that the expert panel accepted the methodology because of the awareness of time pressure, rather than after achieving consensus around the concept of landscape that underlaid it.

4.2 The remarks of the Regional Government

As explained above, the Regional Government set up the terms of reference to develop the classification of landscape units. One of the main constraints implied that the scale of the final product should be 1:25,000. Vallés et al. [12], in their review of landscape mapping exercises in Spain, found that maps at national or regional level were usually produced using scales from 1:200,000 to 1:100,000, while maps produced at province level (NUTS 3) normally used 1:50,000 or 1:25,000, and maps produced at municipal level (LAU 2) usually adopted 1:10,000 (Nomenclature of territorial units for statistics -NUTS- and Local Administrative Units -LAU- are part of a standardized, hierarchical system for dividing up the economic territory of the European Union; for details, see: http://ec.europa.eu/eurostat/web/nuts/overview). There may be well funded doubts regarding whether the goal of producing a 1:25,000 map was finally achieved, and it is probably more likely the resulting map is probably well in the line of those already existing at regional level.

If, according to literature [12], the most appropriate scale for landscape mapping at regional level is 1:200,000 or 1:100,000, it may result confusing why the Regional Government decided for 1:25,000. In our view, this has probably more to do with a sectoral understanding of what is landscape, as officials that develop landscape policies are very often architects and other technicians involved in planning at local scale (physical planning and zoning regulations at local scale very often use scales ranging from 1:500 to 1:10.000). The main implication of the chosen scale (1:25,000) was that the average area of the resulting landscape units remained small enough

Mapping landscape units in Galicia (Spain)   101

(the average unit size in the map described in this work was 1 km²) and that, therefore, the policies that could be applied in the delimited units had to be unavoidably sectorial. For instance, the “low density residential areas” mainly included buildings, transportation infrastructures and some vacant land, all of which is regulated by the sectoral land policies under the label of urbanism. In this way, by choosing the 1:25,000 scale, the Regional Government seemed to create a landscape policy that puts together all sectoral policies, but without integrating them (and therefore without giving them a suitable landscape scale) into an upper hierarchical level. This does not happen without a reason, of course. Sectoral policies have dominated spatial planning in Galicia since the creation of the Regional Government. The 1995 Spatial Planning Act is still underdeveloped after 20 years. The Spatial Planning Guidelines (Directrices de Ordenación do Territorio) have been passed in 2011, but have been watered down as a legal instrument. The landscape units are an instrument included in the Landscape Law of 2008, but the Spatial Planning Act was not modified to include it. Consequently, there is a lack of clarity with regard to the legal hierarchy of both sectoral legal tools and the landscape policy.

5 ConclussionsIn this paper, we presented a methodology for landscape mapping intended to solve a very complex problem in a very short period of time. The methodology has, as a consequence, several drawbacks and some virtues. Regardless of the methodological aspects learnt in the process, the aspects related to the management of the process and the interactions between several different actors seem also very interesting. In particular, the fact that lack of awareness and understanding of other approaches to landscape study may compromise the results of any landscape-related project. The experience presented here stresses the need of multidisciplinary work, of identifying and acknowledging strengths and weakness of each approach when it comes to define concrete and effective landscape policies [25]. And, of course, the need of allowing the appropriate amount of time and resources to achieve that fundamental goal [26].

Landscape policy shall integrate sectoral policies, but sectoral policies cannot give any shape to a landscape policy. Territorial policies seem often kidnapped by sectoral policies. A variety of factors have increased the number of actors involved in the policy process, such as the emergence of the information society, greater

emphasis on public participation and the increasing role of non-governmental organisations, pressure groups and agencies in the decision-making process. All these developments make policy integration increasingly difficult to achieve if enough resources are not dedicated to it.

Acknowledgements: This study was funded by a cooperation agreement signed in 2015 between the University of Santiago de Compostela and the Instituto de Estudos do Territorio (Xunta de Galicia) for the preparation of the Landscape Catalogue of Galicia.

Conflicts of interest: The authors explicitly state that they are not involved in any potential financial, personal or professional conflict of interest regarding the publication of this article.

References[1] Corbelle Rico, E., Crecente Maseda, R., Urbanización, forestación

e abandono. Cambios recentes na paisaxe de Galicia, 1985–2005, Revista Galega de Economía, 2014, 23, 35–52

[2] Brown, G., Brabyn, L., An analysis of the relationships between multiple values and physical landscapes at a regional scale using public participation GIS and landscape character classi-fication, Landscape and Urban Planning, 2012, 107, 317–331

[3] Capotorti, G., Guida, D., Siervo, V., Smiraglia, D., Blasi, C., Ecological classification of land and conservation of biodiversity at the national level: The case of Italy, Biological Conservation, 2012, 147, 174–183

[4] Chuman, R., Romportl, D., Multivariate classification analysis of cultural landscapes: An example from the Czech Republic, Landscape and Urban Planning, 2010, 98, 200–209

[5] Kienast, F., Frick, J., van Strien, M.J., Hunziker, M., The Swiss Landscape Monitoring Program – A comprehensive indicator set to measure landscape change, Ecological Modelling, 2015, 295, 136–150

[6] The Countryside Agency and Scottish Natural Heritage, Landscape Character Assessment. Guidance for England and Scotland, 2002

[7] Wascher, D.M. (Ed.), European Landscape Character Areas. Typologies, Cartography and Indicators for the Assessment of Sustainable Landscapes, Final Project Report as deliverable from the EU’s Accompanying Measure project European Landscape Character Assessment Initiative (ELCAI), Landscape Europe, 2005

[8] Mücher, C.A., Klijn, J.A., Wascher, D.M., Schaminée, J.H., A new European Landscape Classification (LANMAP): A transparent, flexible and user-oriented methodology to distinguish landscapes, Ecological Indicators, 2010, 10, 87–103

[9] Jasiewicz, J., Netzel, P., Stepinski, T.F., Landscape similarity, retrieval, and machine mapping of physiographic units, Geomorphology, 2014, 221, 104–112

[10] Soto, S., Pintó, J., Delineation of natural landscape units for Puerto Rico, Applied Geography, 2010, 30, 720–730

102   E. Corbelle-Rico, J.M. Tubío Sánchez

[11] van Eetvelde, V., Antrop, M., A stepwise multi-scaled landscape typology and characterisation for trans-regional integration, applied on the federal state of Belgium, Landscape and Urban Planning, 2009, 91, 160–170

[12] Vallés, M., Galiana, F., Bru, R., Towards Harmonisation in Landscape Unit Delineation: An Analysis of Spanish Case Studies, Landscape Research, 2013, 38, 329–346

[13] GRASS Development Team, Geographic Resources Analysis Support System (GRASS) Software, Version 7.0. Open Source Geospatial Foundation. USA, 2015

[14] Jasiewicz, J., Netzel, P., Stepinski, T., GeoPAT: A toolbox for pattern-based information retrieval from large geospatial databases, Computers & Geosciences, 2015, 80, 62–73

[15] Goodman, S.N., Fanelli, D., Loannidis, J.P., What does research reproducibility mean?, Science Translational Medicine, 2016, 8, 12

[16] Nosek, B.A. et al., Promoting an open research culture, Science, 2015, 348, 1422–1425

[17] Dirección Xeral de Conservación da Natureza, Plan Director da Rede Natura 2000 de Galicia. Anexo VII. Xeorreferenciación e valoración dos compoñentes da biodiversidade. Consellería de Medio Rural, Xunta de Galicia, 2011

[18] Rodríguez Guitián, M.A., Ramil Rego, P., Clasificaciones climáticas aplicadas a Galicia: revisión desde una perspectiva biogeográfica, Recursos Rurais, 2007, 3, 31–53

[19] Ramil Rego, P., Rodríguez Guitián, M.A., Rubinos Román, M., Ferreiro da Costa, J., Hinojo Sánchez, B., Blanco López, J.M.

et al., La expresión territorial de la biodiversidad. Paisajes y hábitats, Recursos Rurais, 2005, 2, 109–128

[20] Forman, R.T.T., Godron, M., Landscape ecology, John Wiley & Sons, 1986

[21] Cha, S., Comprehensive Survey on Distance/Similarity Measures between Probability Density Functions, International Journal of Mathematical Models and Methods in Applied Sciences, 2007, 1, 300–307

[22] Tress, B., Tress, G., Décamps, H., d’Hauteserre, A.M., Bridging human and natural sciences in landscape research, Landscape and Urban Planning, 2001, 57, 137–141

[23] Price, C., Subjectivity and objectivity in landscape evaluation: an old topic revisited. The economic value of landscape, New York: Routledge, 2012

[24] Antrop, M., A brief history of landscape research, In: Howard, P., Thompson, I., Waterton, E., Atha, M. (Eds.), The Routledge Companion to Landscape Studies, London: Routledge, 2013

[25] Tress, G., Tress, B., Fry, G., Analysis of the barriers to integration in landscape research projects, Land Use Policy, 2007, 24, 374–385

[26] Meijers, E., Stead, D., Policy integration: what does it mean and how can it be achieved? A multi-disciplinary review, Berlin Conference on the Human Dimensions of Global Environmental Change: Greening of Policies-Interlinkages and Policy Integration, Berlin, 2004