Simulations Results & ExpertsViews

Do the agri-environmental policies alter sustainability of rural landscapes?

Jean-Paul Bousset, Nejma Chabab, Geneviève Bigot, Etienne Josien, Teresa Pinto-Correia, Eric Perret, Yves Michelin, Nadine Turpin

Abstract:

“Landscape” means an area, as perceived by people, whose character is the result of the action and interaction of natural and/or human factors (Article 1 – definitions, European Landscape Convention, Firenze 20.X.2000). The definition is clear and nowadays well accepted. Trickier is to progress in the sense of identifying landscape quality objectives, as also demanded in the Convention, and which should depend on the public demand of these landscapes, i.e. their cultural and amenity values.

New EU policies focusing on the development of Europe’s rural areas, including its broad variation of landscapes, and specially the New European Model of Agriculture, which stresses the territorial role of agriculture, increases the need to better understand how these different landscapes are valuated by people, and how farming may contribute to so-called landscape quality. Indeed, people do value landscape patterns, elements, and their combinations (Dramstad et al 2006). They are also able to express their preferences for certain landscape patterns, but there are still challenges to assess these preferences taking care of the diversity and complexity of the landscape (Alkomany 1999), and to combine these preferences with the increasing number of models assessing how policy options can modify the chain "on-farm modifications –> agricultural land use changes –> landscape pattern –> adequacy to people needs and expectations" and reversely how human expectations can be considered in the design of specific policy options can be coordinated with each other.

This paper aims to describe a landscape amenities evaluation tool that aims at filling this gap. This tool is based on the identification of an optimum range of landscape composition, defined through a survey assessment to user groups or expert panels, and its relation, through selected indicators, to the landscape composition resulting from different scenarios.

Keywords : sustainability, policy option, impact assessment, landscape, amenity

1. Introduction
The aim of this paper is to determine which are regional experts' preferences regarding the landscapes in field cropping areas in the Auvergne region, and the cultural and amenity functions these landscapes support. We also want to analyse which factors or elements, in the landscape, determine these preferences. Last, this project aims at contributing to the definition of a method for the evaluation of cultural and amenity functions from expert panels.

In the analysis of sustainability, major attention is usually paid to the economic and environmental aspects. We want to contribute to the development of impact analysis methods that integrate social aspects related to landscapes. In this paper, we relied on the Seamless integrated framework that related bio-physic and economic models at various levels, from field to EU (Van Ittersum et al, 2008), we added a component to assess the consequences of policy options to the amenities provided by a landscape.
The general approach is designed in several steps:
· Identification of the main landscape issues/ units in Auvergne;
· Characterisation of 'field cropping areas' in Auvergne region;
· Identification of the most non productive functions of landscape field cropping areas';
· design of block diagrams that summarize the main issues, characteristics and functions;
· Evaluation, through survey and bloc diagrams (patterns) as visual stimuli, of the characteristics that justify the demand and preferences – by the experts;
· Evaluation of the role of the agricultural systems on these characteristics;
· Evaluation of problems and potentialities for the development of such functions, at farm level.

2. How to involve experts in the assessment of policy impact on landscape amenities?
The framework developed is based on previous projects results and on empirical knowledge of the area cumulated in the team. Joliveau et Michelin (2001), Michelin et Joliveau (2002), Michelin (2000), Joliveau et al (2008), detail the major insights of these results.
The sequential steps considered are:
· Definition of the different types of experts to be considered, in relation to the functions selected;
· Definition of the enquiry scheme to be applied to the experts;
· Definition of the set of patterns ('bloc diagrams' and zooms) to be used as visual stimuli for the assessment of landscape preferences in 'field cropping' areas in Auvergne region;
· Evaluating policy impacts on Landscape Functions (Green Intensification Policy Scenario) in 'field cropping' areas in Auvergne region

The study was conducted having as basis a prototypical municipality which was acknowledged as representative of the 'field cropping areas' in Auvergne region and the landscape units defined, by the experts.

2.1. Expert panel structure and involvement strategy

In our analysis, experts were defined as "representatives of organisations, which are affected by, or which significantly affect, interactions between regional landscapes and 'crop fields' oriented agricultural farming systems, or have skills and knowledge about the issue, or influence implementation instruments relevant to the issue".
Expert panel involved in this work includes experts from 2 different domains: agriculture and environment. Experts are representative people from regional implementation agencies (local governance) or interest groups in Auvergne region. Table 1 presents expertise domains and roles of experts involved.

Table 1: Expertise domains and roles of experts involved in AuvergneCropLand project
Domain Expert Representative for …
Agriculture A1 Interest group
A2 Implementation agency
A3 Interest group
A4 Implementation agency
Environnement E1 Interest group
E2 Interest group
E3 Implementation agency
E4 Implementation agency
E5 Interest group
E6 Implementation agency

The strategy for insuring effective participation of the above experts includes the following stages: i) phone contacts to present the aims and method of the analysis foreseen and to get a date for a face to face interview; ii) face to face interviews with structured semi-opened questionnaires, from which scientists deduce possible consensual positions; iii) group discussion to present the main results of individual interviews – in an anonymous way, and to discuss pre-constructed possible consensual positions.

2.2. Enquiry scheme

2.2.1. Organisation of the survey
The survey was made up of 4 parts:
· Firstly, the interviewees are asked to characterize 'field cropping' areas in Auvergne Region. This question introduces the issue and makes the link with the choice of the field study prototype “Limagne des buttes”. The persons are also asked if they agree with the
field study prototype leading to the visual aids.
· Secondly, a set of questions about the 3 “Block diagrams” are following on. After a ranking proposal based on their first impressions, the interviewees have to point out the differences between the 3 blocks (their answers to this question show how they break down the landscape). For the block put in “first position”, they have to say “who share their point of view and why” in term of space users, they have also to list what they like and dislike on the blocks (this last point underlines the elements making the landscape preferences). They have to define what block is the most interesting as regards several activities (agriculture, tourism, settlements and environment…).
· Thirdly, questions are asked on the 4 zooms of their preferred block: as it was done on the 3 blocks, they have to propose a ranking and to identify the space users sharing their point of view. This step allows the interviewee to precise his ideas on elements that he missed on the larger block.
· Finally, a question about the other landscapes and theirs issues in Auvergne Region make wider the discussion. Then, as conclusion the interviewees can express themselves about the visual aids used and the unreeling of the survey.
Throughout the survey, the interviewees have always been required to make explicit the reasons of their choices and positions

2.2.2. Design of the block diagrams that support the survey
Referring to the FADN universe, 'field cropping areas' – defined as areas of farms belonging to the 'Specialist field crop' FADN farm cluster in Auvergne, which includes" Specialist cereals, oilseed and protein crops", "Specialist root crops", and "Cereals and root crops combined" farming systems, represents a few segment of the area of Auvergne region (5%).
The study was conducted having as basis a prototypical municipality/ area so-called 'Billom like area' as representative area of 'field cropping areas' in Auvergne region.
Three different landscape units (patterns) were used to ask experts on their preferences for the cultural and amenity functions it supports in 'field cropping' areas in Auvergne region;

Each landscape unit (block diagram) is a specific combination of land cover diversity, land cover specialisation and level of intensity in the land use or livestock presence; issues as irrigated fields or orchards and vineyards are considered, as they appear in the Auvergne landscapes. Figure 1 show the three Block Diagrams used to ask experts.
Each of the three ‘Block Diagrams’ is composed of “landscape elements” called “attributes” classified in three categories according to their linking degree with the farm activity: in direct link and in indirect link. The aim is to obtain a catalogue with few attributes which are also functional and simplistic. The other elements on the blocks (road, urban structures and forests…) don't vary with the 3 blocks. They are only there to give a consistent and realistic structure to the picture.
The block-diagrams have been designed according to combinations of three criteria, specialisation of farms, intensity of cropping and diversity of the landscape (see Figure 1).
Each criteria is split in three values, low (1), medium and high (3).
Figure 1: design pattern of the block diagrams

2.2.3. Block diagrams designed for the Auvergne application
The application has been realised on a prototypical municipality, which was acknowledged, by the experts, as being representative of the 'field cropping areas' in Auvergne region and the landscape units defined.
Each of the three block diagrams corresponds to a different evolution state of agriculture.
The BD1 corresponds to the more intensive scenario, the BD2 to an intermediate scenario and the BD3 to the more extensive scenario.
Specialisation Intensity Diversity traduced by a gradient attendance of farm facilities (irrigation material and farm buildings )
traduced by a gradient attendance of areas of ecological regulation (hedges, groves, riverain…)
. The BD1 represents a scheme with a predominance of annual crops, big-sized land areas, with a high attendance of farm facilities like irrigation material and a low rate of areas that ensure ecological regulation (hedges, groves…).
· The BD2 depicts a situation with mixed farming, livestock and diversification crops like vineyards and orchards. The land areas are medium-sized, some farm facilities are remaining and the areas of ecological regulation are more present than in BD1.
· The BD3 represents a scheme with mixed farming and livestock with higher diversification than in BD3. Here, the land areas are smaller and the areas of ecological regulation are even more represented than in BD2.
Figure 2: BD1- predominance of arable crops
Figure 3: BD2 – mix farming
Figure 4: BD3 – diversified mix farming
In addition, each Block Diagram was completed with 4 zooms, which represent 2 different levels of Intensity – Low (1) and High (3), combined with to 2 different levels of Diversity – Low (1) and High (3). Figure 3 shows the 4 Zooms used for the completion of Block Diagram
BD1.
Zoom1 (Intensity = 1; Diversity =1) Zoom2 (Intensity = 1 ; Diversity = 3 )
Zoom1 (Intensity = 3 ; Diversity =1) Zoom2 (Intensity =3 ; Diversity =3 )
Figure 5: Zooms for the completion of Block Diagram BD1

3. Using models to assess the economic and environmental impacts of policy options

3.1. The 'field-cropping area' in Auvergne

The Auvergne region is mainly a low mountain region (600 to 1900 m) with a very high value biodiversity which is preserved with extensive breeding systems (Heckelei T, et al 2005).
Permanent meadows represent 66 % of the agricultural area (Agreste Auvergne, 2004). Main farm orientations are meat cows husbandry (OTEX 44) and milk cows husbandry (OTEX 41) with respectively 28% and 22% of farms, then sheep and goats (OTEX 44) represent 15% of farm orientations. Arable farms are not numerous, less than 10% and crops cover only 42% of the regional arable lands. In this context, current subsidies (1st and 2nd pillars mainly) correspond to 89 % of the regional farm income.
The field cropping area in Auvergne is a 210 000 ha North-South plain, crossed by the Allier river, hemmed by mountains (Monts Dôme and Monts Dore to the West, Monts du Forez to the East) and sprinkled on the edges with low hills. Alluvial or calcareous clay soils or are very deep and fertile. The climate is typically continental and quite dry. 2600 professional farms (average 65 ha) are specialized in cash crops, mainly maize (with an important activity of seeds production with the Limagrain group), soft wheat, sunflower and sugar beet.

3.2. The Green Intensification Policy (GIP) in Auvergne region

Growing concern for the local competition for resource allocation between economic development and natural resource restoration highlights the difficulties of coordinating agricultural and environmental policies at EU, national and regional levels. More precisely,
local policy makers have to face the general evolution of economic forces (like prices changes, increased competition between countries for labour and resources use), they have to coordinate their action with EU wide policies, and operate through specific regional stakes.
The GIP scenario explores the ways a regional agro-environmental policy can improve the sustainability of the farming systems in a region, in a context of general food price increase and related environmental threats.
We assume in this scenario that the level of increase for agricultural products prices and for production costs modifies the relative profitability of agricultural production over Europe.
Even in mountainous areas farmers are liable to intensify their production systems. As the main stakes for the Auvergne region are the preservation of the high level of biodiversity, the natural landscapes and the water quality with a stabilization of the rural population, local stakeholders would like to compare different policies at the local level.
The GIP scenario assumes that : (i) prices of cattle and cropping products are increased about x % relatively to the baseline (x varying between 25% to 100%); (ii) milk quotas and set-aside obligation are removed; (iii) farmers develop intensive technologies based on the use of more machineries, fertilizers and pesticides. GIP consists of – for the regional government – attempting to protect the environment in circulating an incite premium (PREM varying between 10 to 50 Euros per ha UAA) to farmers that limit the nitrogen balance on their farm (threshold = 50 …100 kg / ha).

3.3. A bio-economic model
FSSIM-MP (Janssen et al., 2008) is a comparative static mathematical programming model with a non-linear objective function representing farmer’s behaviour. It is an individual farm model calibrated at the farm level and working with exogenous prices coming from different sources. The principal FSSIM-MP specifications are: (i) a static model with a limited number of variants depending on the farm types and the conditions to be simulated. Nevertheless, to incorporate some temporal effects, agricultural activities are defined as “crop rotations” and “dressed animal ” instead of individual crops and animals; (ii) a risk programming model based on the Mean-Standard deviation method in which expected utility is defined under expected income and risk (Hazell and Norton 1986); (iii) a positive model, where the main objective is to reproduce the observed production situation as precisely as possible by making use of the observed behaviour of economic agents and; (vi) a generic model designed with the aim to be easily applied to different regions and conditions.
The mathematical structure of FSSIM-MP is formulated as follows:
Maximise: U = P’x + S’x – d’x - x’Qx/ 2-K-
Subject to: Ax B; x0
Where: U is the variable to be maximised (i.e. utility), P is a (n x 1) vector of gross margin of each agricultural activity, S is a (n x 1) vector of subsidies per unit for each agricultural activity (depending on the Common Market Organisations (CMOs)), d is a (n x 1) vector of parameters of the cost function, Q is a (n x n) symmetric, positive (semi-) definite matrix of the cost function (the estimation of the vector d and the matrix Q depends on the calibration approaches), x is a (n x 1) vector of the level of agricultural activities, K represents single fixed costs (including annuity for investment) at farm level, A is a (m x n) matrix of technical coefficients, B is a (m x 1) vector of available resource levels, is a scalar for the risk aversion coefficient, is the standard deviation of income according to states of nature defined under two different sources of variation: yield (due to climatic conditions) and prices.
The agricultural activities (i) are defined in FSSIM-MP as a combination of crop rotation (r), soil type (or agri-environmental zone) (s), production technique (t) and production orientation (sys) (i.e. i @ r,s,t,sys). That is, an agricultural activity is a way of growing a rotation taking into account the management type. However, if data on crop rotations are missing the agricultural activities can be defined using individual crops (i.e. mono-crop rotations).
The principal technical and socio-economic constraints that are implemented in FSSIM-MP are: arable land per soil type (or agri-environmental zone), irrigable land per soil type, labour and water constraints. The same rule was applied for all of these constraints: the sum of the requirements for each resource cannot exceed resource availability.
FSSIM-MP can be calibrated using any or all of the following approaches, depending on the application type: (i) the risk approach, (ii) Monte Carlo, (iii) the standard Positive Mathematical Programming (PMP) procedure (Howitt, 1995), (iv) the Rhöm and Dabbert’s
PMP approach (Röhm and Dabbert, 2003).

1 The concept of ‘dressed animal’ represents an adult animal and young stock taking into account the replacement rate.
Figure 6: FSSIM-MP structure (adapted from Janssen et al., 2008)
FSSIM-MP has a modular set-up, including modules on crops, livestock, perennial, investment, premium, risk, policy and Positive Mathematical Programming (PMP). These modules are linked indirectly by an integrative module named the “common module”
involving the objective function and the common constraints (Figure 2.1). Each module includes two GAMS files. The first one links the data-definition and the module’s equations and the second one contains the module’s equations. Each module generates at least one variable which is used to define the common module’s equations, thus providing a link between the different modules.

Policy impacts on Landcape Functions (Green Intensification Policy Scenario) in 'field cropping' areas in Auvergne region have been evaluated from the results of FSSIM simulations in a virtual (prototypical) farm so called FT4 farm-type in SEAMLESS Database,
whose the features (size, crop-rotation, practices and economic results…) are the average value of features of the actual FADN farms belonging to the 'Specialist field crop' cluster in FADN universe in Auvergne.

4. Assessment of amenities provision at landscape level (links of the two approaches)

Assessment of landscape amenities is realised using a set of indicators that depict the following functions:
- diversity is an environmental indicator accounting for land cover diversity (and, conversely, land cover dominance, considering that a low diversity conversely means the dominance of few land cover classes) – relevant for biodiversity and for environmental quality, in relation to cropping pattern and concentration/distribution.
The indicator can also be used as social indicator, as it can be a measure of landscape General Algebraic Modelling System which is used to program the model diversity (IRENA considers for Landscape State: presence of crops + linear elements + patch density), and thus an indicator of landscape interest for several noncommodity functions, as hunting, recreation and 2nd housing. In our case study, is characterised by the quantity and the variety of factors of ecological regulation as hedges, groves, banks….

The more the variety of crop rotation system is high, the more there are numerous areas of ecological regulation; we can even consider that sometimes the plots of land become areas of regulation (in the case of the grassed vineyards or orchards).
- specialisation is an environmental indicator which indicates the dominant type of farm in a region, but also the dominant agricultural land use (type). It indicates the quality and proportion of the dominant land cover class at landscape level, so that knowing what is the diversity of the land cover classes, at the same time it is possible to know what is the dominant one and of what quality. Specialisation fits with the direct attributes: land cover (crops, vineyards, orchards or meadows) and land area size (small, medium or big-sized). For example, large-sized land areas, few meadows and no diversification (no vineyard and orchards) define the Block Diagram 'BD1', which has the highest degree of specialisation.
- intensity is depicted as the area of agriculture under intensive use, including the presence of farm facilities. It is characterised on blocks by the attributes in indirect link (farm facilities and areas of ecological regulation). From the model outputs, the criterion refers to the intensity of the farming practices.

5. Crossing the Expert Views and the value of the landscape functions

So far, we have two values for each function, one is expert-made, and the other is assessed based on indicators. There is a strong need to relate them, considering that they do not have the same significance, the same grounding and reference levels. We used the IFS, the Index of Functions Suitability to the landscape, to describe the gap between expert values and indicator values. The IFS describes the inverse of the difference between the expert designed values and the amenity values that result from the scenario – and which is based on the indicators. So the higher the IFS, the best the scenario is for that landscape pattern. In this way it is possible to evaluate the display of selected functions according to future possible scenarios.

5.1. Expert views

5.1.1. Experts' views on 'Billom like area' as representative area of 'field cropping areas' in Auvergne

All experts have acknowledged the used 'Billom like area' as representative of the 'field cropping areas' in Auvergne region.

Table 1 reports individual comments of experts Domain Expert Representative for …
Agriculture

A1 Interest group Yes - Other possible choice: 'Llimagne noire'
A2 Implementation agency Yes Limagne noire = no landscape issues
A3 Interest group Yes Limagne noire = no landscape issues
A4 Implementation agency Yes - Other possible choice: 'Llimagne noire'

Environnement

E1 Interest group Yes + Actual landscape issue
E2 Interest group Yes + Exemple of 'hilly' civilisation
E3 Implementation agency Yes Like in Devez (43)
E4 Implementation agency Yes + Limagne noire = no landscape issues
E5 Interest group Yes + Exemple of Landscape diversity issue
E6 Implementation agency Yes + Exemple of biodiversity issue

5.1.2. Experts' views on the main differentiation factors of the used landscape patterns (Block Diagrams)

Most of the experts pointed out Division, Erosion protection (hedges, trees), Intensification (irrigation, silos), Diversification (number of different crops, orchards) as differentiation factors among the three block diagrams. Most of the experts agreed to see BD3 as the most suitable landscape pattern to optimise environmental and residential functions in the modelled field cropping areas. On the other hand, the expert views varied a lot on the most suitable landscape pattern to optimise economic functions in such area.

5.1.3. Experts' preferences for the landscape patterns and the reasons/explanations for their preferences

Sixty percent of the experts selected Block Diagram BD2 as preferred landscape pattern for the 'field cropping areas' in Auvergne region. The most frequent explanation of their preference is division (number of areas of ecological regulation), followed by diversification (numerous farming system productions), and erosion protection. The other experts (40%) selected Block Diagram BD3 as preferred landscape pattern for the 'field cropping areas' in Auvergne region. The most frequent explanation of their preference is biodiversity (number of areas of ecological regulation), followed by diversification (link to proximity markets),
erosion protection and farm density in the area. Any expert selected BD1 as preferred landscape pattern, but 20% of the experts ranked such a Block Diagram as a challenger of the best one. BD1 was rejected for the few level of division of the area, issue of erosion linked to strong slopes in the upper part of Block Diagram 1.

It is now possible to compare the provision of landscape amenities from the two approaches.
5.3.1. Calculating the IFS with an optimum value like a range of values

5.3.2. Calculating the IFS with an optimum value like a unique value

6. Conclusion

Survey of expert panel concerning landscape scenarios in Auvergne Region demonstrated the efficiency of the applied an approach based on the use of Block Diagrams as landscape patterns. The survey was conducted having as basis a prototypical municipality which was acknowledged as representative of the 'field cropping areas' in Auvergne region.

Such a methodology has allowed us to evaluate the preferences of regional experts in relation to the landscape in field cropping areas in Auvergne region, for the amenity functions it supports, as well as factors and elements, in the landscape, which determine these preferences.

Expert discussion demonstrated the need of skills and knowledge from at least two different domains: agriculture and environment. In addition, these discussions underlined the need to ask specialists in specific environmental domains as erosion; ecological regulation (e.g.hedges…)

Block Diagram (landscape patterns) demonstrated their ability to represent the main landscape issues and landscape units in Auvergne region. Although they are simplifications of actual landscape features, they have been acknowledged by experts as effective visual stimuli to reveal the value system and preferences of each expert, which was acknowledged as strongly subjective.

They demonstrated their ability to represent the most non productive functions of landscape in ‘field cropping areas' –
(division, erosion protection, biodiversity via ecological regulation (rive-rain…)), as well as different combinations between them and agricultural systems.

Moreover, expert discussion from/ on block diagrams have allowed us to evaluate problems and potentialities for the development of such functions, at local level.

Sixty percent of the experts selected Block Diagram 2 with medium Land cover specialization, medium Land cover diversity and medium intensity as preferred landscape pattern for the 'field cropping areas' in Auvergne region. The most frequent explanation of
their preference was division, followed by diversification and erosion protection. The other experts (40%) selected Block Diagram 3 with low Land cover specialization, high Land cover diversity and low intensity as preferred landscape pattern for the 'field cropping areas' in Auvergne region. The most frequent explanation of their preference was biodiversity, followed by diversification and erosion protection. Any expert selected Block Diagram with high Land cover specialization, low Land cover diversity and high intensity as preferred landscape pattern, but 20% of the experts ranked such a Block Diagram as a challenger of the best one.

Expert discussion indicated that a consensual scenario should include the upper part of Block  Diagram 2 and some elements of the down part of Block Diagram 3, that is to say, for example: adding areas of ecological regulation to the lower part of BD2 (rive-rain, hedges…); adding orchards, vineyards in the upper part of BD2.

Calculating the IFS with an optimum value like a unique value showed a gap equal to 0.33 between the most expressed expert preferences and the tested policy scenario in the 'field cropping areas' in Auvergne region.

References
Al-Kodmany, K., 1999. Using visualisation techniques for enhancing public participation in planning
and design: process, implementation and evaluation, Landscape and Urban Planning, 45: 37-45.
Dramstad, W. E.; Tveit, M. S.; Fjellstad, W. J.; Fry, G. L. A., 2006. Relationships between visual
landscape preferences and map-based indicators of landscape strutuer, Landscape and Urban
Planning, 78: 465-474
Howitt, R., 1995. Positive Mathematical Programming. American Journal of Agricultural Economics
77, 329-342.
Janssen, S., M.K. Van Ittersum, K. Louhichi., P. Zander, N. Borkowski, A. Kanellopoulos, H.
Hengsdijk et al. (2007). ‘Integration of all FSSIM components within SEAMLESS-IF and a stand
alone Graphical User Interface for FSSIM’, D3.3.12, SEAMLESS integrated project, EU 6th
Framework Programme, contract no. 010036-2), www.SEAMLESS-IP.org, 53 pp.
Joliveau T., Michelin Y., 2001. Modèle d'analyse et représentation pour la prospective paysagère
concertée: deux exemples en zone rurale In : Représentations spatiales et développement territorial
(Edit : S. Lardon, P. Maurel, V. Piveteau), Paris, Hermes:239-266
Joliveau T., MichelinY., Ballester P., 2008 : chapitre 11 ; éléments et méthodes pour une médiation
paysagère in « paysage et information géographique » sous la direction de T. Brossard et J.C.
Wieber, Paris, Hermes, lavoisier : 257-286
Michelin Y., 2000. Le bloc-diagramme : une clé de compréhension des représentations du paysage
chez les agriculteurs ? Mise au point d’une méthode d’enquête préalable à une gestion du paysage
en Artense (Massif central français). Cybergéo, 118, 12p
Michelin Y., Joliveau T., Breuil J., Vigouroux L., 2002: Le paysage dans un projet de territoire,
démarche et méthode expérimentées en Limousin, Limoges, Chambre d'agriculture de Haute
Vienne, 66 p
Pinto-Correia T., Machado C., Picchi P., Alkan Ollson J., Turpin N., Bousset J.P., Michelin Y.,
Chabab N., Bockstaller C., Bezlepkina I., 2009, Landscape amenity model, in in SEAMLESS-IF,
PD1.2.1, SEAMLESS integrated project, EU 6th Framework Programme, contract no. 010036-2,
www.SEAMLESS-IP.org , 53 pp.
Röhm, O., Dabbert, S., 2003. Integrating Agri-Environmental Programs into Regional Production
Models: An Extension of Positive Mathematical Programming. American Journal of Agricultural
Economics 85, 254-265.
Van Ittersum, Martin K., Ewert, Frank, Heckelei, Thomas, Wery, Jacques, Alkan Olsson,
Johanna et al., 2008. Integrated assessment of agricultural systems - A component-based
framework for the European Union (SEAMLESS). Agricultural Systems, 96(1-3), 150-
165.