Int. J. Global Environmental Issues, Vol. 5, Nos. 3/4, 2005

Graphic tools for data representation in integrated analysis of farming systems Tiziano Gomiero
c/o Dr. Maurizio G. Paoletti – Department of Biology, Padua University, Italy E-mail: tiziano.gomiero@libero.it

Mario Giampietro*
Istituto Nazionale di Ricerca per gli Alimenti e la Nutrizione, Roma, Italy E-mail: giampietro@inran.it *Corresponding author
Abstract: This paper provides an overview of graphic tools that are currently used for representing integrated analysis of farming systems and environmental systems. The overview includes also an appraisal of the various methods in terms of pros and cons. Some key issues concerning how to implement a sound integrated graphical representation are also discussed. We believe that the usefulness of sound graphical representations lies in helping social debate and rising awareness among stakeholders about critical trade-offs and the complex nature of living systems, as well as in forcing both researchers and stakeholders to make clear to themselves and to the others the process by which they represent their own reality. Keywords: farming systems; environmental systems; integrated analysis; integrated representation; multiple criteria; graphics tools; social participation. Reference to this paper should be made as follows: Gomiero, T. and Giampietro, M. (2005) ‘Graphic tools for data representation in integrated analysis of farming systems’, Int. J. Global Environmental Issues, Vol. 5, Nos. 3/4, pp.264–301. Biographical notes: Tiziano Gomiero after graduating in Nature Science from the Padua University (Italy), he obtained a Master in Ecological Economics and Environmental Management and then the PhD in Environmental Science from the Universitat Autonoma de Barcelona (Spain) with a thesis concerning the development of models for integrated analysis of farming systems (under the direction of Dr. Mario Giampietro). He worked in international research projects concerning rural development and integrated farming system analysis both in Asia and Europe. At present he is carrying on projects concerning environmental management, environmental education and organic agriculture for Italian Institutions. Mario Giampietro (PhD Integrated analysis of Food Systems) is the Director of the Unit of Technological Assessment at the National Institute of Research on Food and Nutrition (INRAN). Professor Giampietro’s expertise covers issues such as multi-scale integrated analysis (MSIA), biophysical analysis of agricultural and socio-economic development, agro-ecosystem analysis with specific experience on sustainability research in Southeast Asia and China. He has been Visiting Professor at the Cornell University, Autonomous Copyright © 2005 Inderscience Enterprises Ltd.

Graphic tools for data representation in integrated analysis
University of Barcelona, and Visiting Scholar at the University of Wageningen, and University of Wisconsin-Madison. He has been the organiser of the Biennial International Workshop of Advances in Energy Studies held, since 1996, in Portovenere, Italy.



The usefulness of integrated approach in farming system and environmental analysis
What is to be sought in designs for the display of information is the clear portrayal of complexity. Not the complication of the simple; rather the task of the designer is to give visual access to the subtle and the difficult – that is, the revelation of the complex. Edward R. Tufte (1984, p.191)

In this paper, we present a broad overview of important graphical representations found in literature in relation to integrated (concerning multiple criteria) analysis of farming and environmental systems. The graphic tolls dealt with in this paper represent tools aiming at easily convening (complex) information to the stakeholders, information coming from different domains (e.g., environmental, economic, social). This to help stakeholders (among them being policy makers, technicians, farmers, or lay people) to interact and understand each other perspectives, and to discuss about the trade-offs implied in the alternative policies in relation to the changes that they wish to make. It should be noted that different to the many available optimisation methods, based on more or less complicated mathematical algorithms (see for instance European Operational Research Societies – http://www.euro-online.org website, or papers published in journals in the field of operational research), these tools do not claim to provide ‘optimal solutions’. These methods are characterised by recognising the impossibility to aggregate incomparable and incommensurable data, so that they keep them separated and linked to their own domain. The contrary resulting in a single figure, which tells very little about the behaviour of the system under analysis, and offers even less room for discussion. It is not the aim of the paper to go into details on this issue. However, we wish to point out that comparability and commensurability of values are at the core of neoclassical economic approach, even concerning environmental evaluation, while the concepts of incomparability and incommensurability of values are at the core of the field of ecological economics (Funtowicz and Ravetz, 1990; Cabeza Gutés, 1996; Munda, 1997, 2004; Martinez-Alier et al., 1998; Giampietro, 2004).

1.1 The complex nature of farming systems requires integrated (considering multiple criteria) approach
The complex nature of farming and environmental systems makes integrated approaches necessary, both in order to improve their understanding and in turn in the implementation of sound management strategies. Complexity derives from the many different components involved, each with the property to self-organise and evolve in time. This being the very same nature of nature of living systems, among them being


T. Gomiero and M. Giampietro

individuals, species, ecosystems, or social systems (there is an extensive literature available in this field for review, see for instance: Wiener, 1948; Simon, 1962; Odum, 1971; Prigogine, 1980; Allen and Starr, 1982; Rosen, 1991, 2000; Kauffman, 1993; Lewin, 1993; Holland, 1995; Wolf and Allen, 1995; Bak, 1996; Tainter, 1996; Levin, 1998; Kay et al., 1999; Holling, 2001; Gunderson and Holling, 2002; Giampietro, 2004). Taking as an example the case of farming system: a biophysical scaling up leads us to move from soil (with its characteristics), to its fauna (micro-organisms whose activities are essential for nutrients cycling), to crop species (and the cultured pattern). When adopting a socio-economic perspective, we can start from the farmer, and moving up to the household, the rural community in which the household lives, to arrive at local, regional, and national socio-economic systems. Using an ecological perspective, we can have: local agro-ecosystems, the watershed, up to the natural mechanisms generating biogeochemical cycles of water and nutrients that stabilise and constrain the boundary conditions of the farming activity. An adequate representation of a productive system requires a multi-dimensional, or multi-criteria, approach, where many perspectives and level of analysis have to be taken into account (e.g., economic, environmental, social, cultural), and many stakeholders (e.g., farmers, consumers, citizens) involved in the decisional process (Prigogine, 1980; Altieri, 19871 Conway, 1987; Beets, 1990; Pretty, 1995; Wolf and Allen, 1995; Giampietro, 2004; Chambers, 1997; McConnell and Dillon, 1997; Bland, 1999; Gliessman, 2000, 2001; Hall et al., 2000; Dixon et al., 2001; McCown, 2002). Indicators used in economy and ecology are both qualitative and quantitative and then multi-criteria algorithmic (mathematical) methods able to deal with mixed information are considered very useful such as PROMETHEE (Brans et al., 1986; Brans and Marescal, 2002; Brans and Mareschal, 2004), see also the dedicated website http://www.promethee-gaia.com), NAIADE (Munda, 1995), REGIME, and ELECTRE (for details see for instance Keeney and Raiffa, 1976; Bana e Costa, 1990; Munda et al., 1994; Beinat and Nijkamp, 1998; Janssen and Munda, 1999; Goitouni and Martel, 1998; Nijkamp and Ouwersloot, 1998; Nijkamp and Vreeker, 2000; Janssen, 2001; Belton and Steward, 2002; NERA, 2002). According to Munda (1993, 1995, 2004), multi-criteria methods that try to take mixed information (qualitative and quantitative) into account face with two main problems: • • the problem of equivalence of the procedures used in standardising the various evaluations of the performance of alternatives according to different criteria the problem related to the available information as it concerns the uncertainty (stochastic and/or fuzzy) contained in this information.

Therefore, the combination of different levels of measurement with different types of uncertainty has to be considered as an important research issue in multi-criteria evaluation. These methods, however, are based on matrixes analysis and complex mathematical algorithms, often aiming at reaching optimal solutions. The graphical representations presented in this paper differ from multi-criteria mathematical approach by limiting themselves to the representation (in a more of less detailed way) of the value taken by the indicators within some viability or feasibility domain. They are not optimisation methods, nor claim to offer the best solution. They are characterised by being easy methods to presenting the trade-offs

1985. whatever we do. we are performing a manipulation. 1989). A number says very little if it is not accompanied by a detailed explanation of the assumptions and process through which it has been generated (Box. This unavoidable presence of arbitrariness is amplified by the fact that any graphical representation of a given problem or situation – when considering different dimensions and multiple criteria – does require a lot of work on the original set of data to guarantee an overall coherence and clarity in the final image (Tufte. 1983. 2004). allowing us to summarise the general behaviour and to study detail. 1983. Kosslyn. but can also be used to hide on purpose some relevant aspects of it. Cleveland. When organising various sets of data within a graph. 1988. 1976. (b) The definition of the identities of (=the set of relevant attributes and observable qualities associated with) the elements to be included in a model. This is why a discussion on procedures to be followed to generate sound and effective graphical representations deserves the maximum attention (Tufte. Cleveland. Funtowicz and Ravetz. Living in a world dominated by the influence of media. in fact. one of the major advantages provided by graphical representation is that they can show the presence of patterns that would be difficult to perceive in a matrix of alpha-numerical data (if the same selection of sets of data were presented in the form of conventional tables). . When dealing with various sets of data. Zar. Larkin and Simon. A manipulation that will increase the degree of arbitrariness implied by the chain of decisions required to select a particular realisation of how to perceive and represent a complex problem. Tversky and Kahenman. Cleveland and McGill. Cleveland and McGill. Cleveland and McGill. 1985. this very same special ability can generate problems. Tufte. Graphical methods tend to show data sets as a whole. just the last act of a long process that started with: (a) The definition of the goals of the study. Graphical representations have the ability to suggest an overall ‘meaning’. 1990. 1985. Tufte. Even the representation of simple ‘numbers’ is not immune from such a problem. 1984. 1985). the final decision let to social agreement and social responsibility. 1987. pictures not only can make easier the comprehension of a situation.Graphic tools for data representation in integrated analysis 267 involved in the decisional process. 1985. That is. Cleveland. 1983). 1981. 1983. we are all very aware of how the maker of a picture can affect the perception of a given situation that it is conveyed to the observer by that picture. 2004. 1985). (c) The selection of proxies (=variables) that can be used to encode changes in relevant characteristics of the relevant identities considered in the problem structuring. and therefore induce a biased perception of the reality (Tufte. Gomiero. 1985. 1990. Cleveland and McGill. 1969. The writing of a set of data to be used as indicators is. 1983. 1990. 1985). Giampietro. 1985. Cleveland. a sort of gestalt emerging from of the data set (Cleveland. 2 The usefulness of graphic representation for integrated analysis Graphs and graphical representations provide a powerful tool for analysing scientific data as well as they are useful to quickly and easily convey data and information to the reader (Simon. 1990. This definition determines the scale adopted in the model. On the other hand.

” quoted in Smil. forces awareness in the analyst (Simon. To do that. 3. moral. …) of the various alternatives or different realisations of that system typology. Giampietro (d) The setting of a measurement scheme making possible to gather data. which should be assigned to perceptions and representations of the reality over different descriptive domains. Gomiero and M. we can no longer expect to obtain a simple and unique ordinal ranking (1. steps (d) and (e) can be replaced by the use of second hand data. Each one of these steps entails possible sources of confusion among those using the final model about the shared meaning. that is to say to create a cross-contextualisation among different scales and criteria (e. non-equality implies that one is greater than the other. This step. basic assumptions) have been used to build the structured information space carrying the integrated representation.g. technical) into a single definition of quality. • 3 From the single number to ‘pattern’ representation As stated by Jacquard: “When we are comparing one number to another. which introduces a new problem of comparability of the quality of these second hand data. (e) The actual processes of measuring. This would require being able to give a profile of different . being the last of a long chain. it implies that they are different. Enhancing the robustness of the characterisation of various alternatives (or compared situations). therefore.g. where. but in relation to different dimensions).. at one given scale. we should be able to reduce the heterogeneous information carried out by different indicators referring to non-equivalent criteria (e. p. ecological. by covering the same criteria using indicators referring to different scales or representing the same element. 1988). and why the various external referents (direct data.268 T. 2. when. 1993. aesthetic. Alternatively. an indicator acquires a strong identity as it maps on many different domains. implies adding another step to this chain: (f) Organising the data in a graphical representation. when performed carefully. This can be obtained by looking always for a ‘mosaic effect’ in the set of indicators used. economic. indirect data. when we are comparing sets. In this way.. Organising data in a graphical representation.32) When using different sets of data referring to non-equivalent criteria (when accounting for the relevant characteristics of a given system typology). The overall process. This can be obtained by making as explicit as possible. This is not a plea motivated by moralistic considerations. it is a statement of logical fact. This is why one should be aware that during the process of building a graphical representation of an integrated analysis the scientists must always have two goals clear in mind: • Making aware as much as possible the reader of potential misunderstandings in the transmission of information. cannot be performed without an explicit knowledge and linkage to the previous ones. This would require defining an ultimate ‘absolute’ index of performance for the typology of system.

we would like to make an important distinction about possible ways in which multiple data can be handled either for a process of decision making or for a simple comparison. have been proposed by Munda. the analysts assume that it is possible to deal in substantive terms (in general terms and with no possible contestations) with: • • technical incommensurability (it is to say that it is impossible to reduce to a single model analysis referring to non-equivalent descriptive domains) social incommensurability (it is to say that it is normal to find legitimate but contrasting views in social actors about what should be considered as an improvement). 1996. therefore. on this point see also Giampietro. per year (Pearce and Turner. In the first example. A system typology is required to have reference against which it becomes possible to evaluate alternatives or to compare different systems. This is why a multi-criteria evaluation requires addressing explicitly ‘values’. and indicators. technical incommensurability and social incommensurability. The problem is that such a reduction and comparison can only be done in a give point in space and time (in a special situation) by a given group of social actors. sex. Such an index then is assumed to provide a reliable measure for the overall system performance. TEV expressed in US$ of a given year. Tisdell. and personal aspirations. The issue of sustainability implies handling indicators and data referring to different scales and dimensions of analysis. (The two terms. social status. criteria. This heterogeneous information space can be processed using different rationales.Graphic tools for data representation in integrated analysis 269 weights – the relative importance they have in a given analysis – to the set of considered indicators. religion) operating in different places and in different moments of their life trajectory all possible situations in which the considered analysis can be relevant for action. A process of multi-criteria analysis. the . 2004. Put in another way. Those following this approach propose protocols. which are used to aggregate a set of indicators referring to different criteria into a single numerical index. 2001). 1993). which are reflected in preferences. the GNP is corrected to account for the effect – expressed in monetary value – of changes in the environment. we cannot expect to use the same standard algorithmic protocol to weight incommensurable criteria. implies to choose among different combinations of incommensurable typologies of ‘pros’ and ‘cons’ when defining objectives. Others examples of the same idea are the Sustainability barometer (Prescot-Allen. and the ISEW developed by Daly and Cobb (1989). cultural identity. To put better in perspective the discussion about the role of graphical representation in a process of multi-criteria evaluation. different indicators referring to different dimensions of sustainability are collapsed into a single numerical index. three approaches are relevant for our discussion: Aggregation of indicators referring to different dimensions into a single numerical index In this approach. In particular. when considering: • • all possible social agents (of different age. Also in these two examples. It is impossible to even think that it would be possible to do that in general terms. In this approach. culture. An example of such an approach is the Total economic value in environmental economics. 2004). 1990.

This approach requires the following inputs: • • • • a set of relevant criteria used to evaluate the performance in relation to the relevant objectives associated to the analysis a set of attributes and indicators for each criterion to characterise the performance a set of possible options (or the given set of systems to be compared) a profile of weighting factors associated to the attributes of performance determined in the previous steps. by having available these four inputs. Algorithmic multi-criteria evaluation (ranking of different alternatives using: a given impact matrix... 1998. a given profile of weighting factors and a given algorithm) In this approach. Bana e Costa. Cartesian axes) has the only goal to improve the exchange of information among those participating in the process.g. ISEW. as well as in dedicated journals (e. it becomes possible to use algorithms to process the information space organised in this way and to generate a ranking among the set of alternatives (or systems to be compared). the graphical organisation of data (e. which is the object of scrutiny.. the very choice of how to organise the representation of relevant issues is itself a step. In this situation. US$ of 1987). 2001. 1994. 1998. Gomiero and M.g. technical incommensurability is avoided by not aggregating different indicators into a single number. 2002. 1990. it becomes possible to deal with the challenge implied by technical and social incommensurability. again it is a number based on monetary value (e. Belton and Steward. When accepting as valid these assumptions. A description of tools and procedures adopted in multi-criteria analysis is available in many books and review papers (e. Within this approach. Goitouni and Martel. Beinat and Nijkamp.g. the social multi-criteria evaluation has to be based on two processes having two distinct goals: . Giampietro sustainability barometer. 2002)..270 T. Social multi-criteria evaluation process (generation of a representation of the issues to a participatory process of integrated assessment of alternatives) The term social multi-criteria evaluation has been proposed by Munda (2004) to explicitly acknowledge a systemic impasse found when attempting to apply multi-criteria analysis to problems of sustainability characterised by high levels of technical and social incommensurability. 1986. The basic rationale associated to this approach is that. Actually. the various options (or the various systems to be compared) are characterised in a multi-criteria framework. triangles. Journal of Multi-criteria Decision Analysis. Janssen. it could be more useful to keep separated the two processes. in the second example. rather than attempting to collapse the descriptive and the normative side into a single process of aggregation. European Journal of Operational Research). Munda et al. it is qualitative index.g. When operating within this rationale. Brans et al. and on the normative side social incommensurability is dealt with by a determination of an agreed upon profile of weighting factors. in the form of radars. NERA.. On the descriptive side..

. Spash and Clayton.. • 4 A survey on integrated graphical representations Listing early examples of graphical integrated representations linked to sustainability Gallopin 1996 describes the work of: (a) Dansereau (1971. 1996) – a star diagram (spider web). 1997).. 1996). for a detailed account of the methods here presented and applications to real case study. 1999. The goal of the analysis is that of providing an integrated assessment of the most pressing needs of African population. 1977 in Gallopin. 2000. Masera and López-Ridaura.g. fears.. Because of the limit of space of this paper.. goals. Vreeker et al. The examples of graphical representation discussed in this paper are: • • • • • • • • • • • AMOEBA: which is a Dutch acronym used for a special form of radar diagram sustainability barometer (SB) sustainability reference system (SRS): kite diagram sustainability assessment map (SAM) prototyping integrated and ecological arable farming system (I/EAFS) intervention impact assessment (IIA) mixing triangle kite diagram for NUSAP applications pie for policy performance index (PPI) the flag model multi-objective integrated representation (MOIR).Graphic tools for data representation in integrated analysis • 271 on the descriptive side: guaranteeing quality in the activities aimed at handling the heterogeneous information space required for perceiving and representing a problem on different scales and dimensions on the normative side: guaranteeing quality in the activities aimed at handling the heterogeneous universe of values.g. 2002). aspirations found in the universe of different social actors relevant for sustainability. the goal of the analysis is a qualitative assessment of environmental health. 1996. Graphical integrated representations range from the use of Cartesian plain (e. In this case. Gallopin. 2001) to simple radar diagrams (called also spider web in market research) to represent multiple indicators of sustainability (e. Additional details for some of these approaches are available also in internet. for these we make reference to the relative website. (b) Bugnicourt (1979 in Gallopin. Clayton and Radcliffe. to quite complex figures based on radar (or other shape) diagram (e. López-Ridaura et al.g. who adopt a similar approach in order to provide an integrated assessment of the most pressing needs of African population. Bossel. 1996. we refer the reader to bibliography section. 1996. divided into a number of sectors corresponding each to an environment component. Prescot-Allen.. 2001.

. Figure 1(a) AMOEBA representation (coastal monitoring: amoeba of the Aegean sea.. abundance of organisms belonging to key species) to monitor the state of the ecosystem at different scales – indicators are clustered for typologies of species operating in different spatial domains and on different scales the actual state is put in relation to a reference situation.. LEEC. 1995. 1992. ten Brink. referring to characteristics and events occurring in different places and scales. 1997. Tonon et al. 2000.. 2002). by the Ministry of Transport. 1999) The acronym AMOEBA stands for: A General Method Of Ecological and Biological Assessment (but it is based on a corresponding set of Dutch words). Heyer et al.. The main relevant features are: • different indicators (e. 1981.g. (1981) for providing an integrated description and assessment of aquatic ecosystems (Figure 1(a)). Bell and Morse. de Zwart and Trivedi. after Verhagen. . 1993.1 The AMOEBA approach The AMOEBA approach is becoming a popular graphical representation in the field of integrated analysis of sustainability (e.3 Wefering et al. 2000. linked to the ability to put in perspective the information given by the indicators in relation to expected values makes this graphical method suitable for structuring the information used in the process of decision making.g. in the Netherlands. ten Brink et al. Gomiero and M. 2001.272 T. 1999. Public Works and Water Management in the framework of the 3rd National Water Management Policy Plan.2 Bockstaller et al.. Masera and López-Ridaura. 1999. The goal of this method is that of providing an integrated view of the ecological situation of a given environmental system. • The ability to handle different sets of indicators.. Giampietro 4. This graphical presentation technique has been developed by ten Brink et al. 2003. 2000. Bockstaler and Girardin. Sucur. López-Ridaura et al. 2000. Verhagen..

They all lie at the same distance from the origin and therefore represent a circumference of a circle used as benchmark. they will. Comments Although ten Brink (1992. That is. p. various indicators can be represented on the axes moving out from the origin on a standard scale from 0 to 100. see for instance Sucur (1993). it is immediately clear which values of the various indicators (characterising the actual state of the monitored ecosystem) fall short or exceed the target. and López-Ridaura et al. reference numbers and actual numbers are given for a number of key species. An alternative representation using the same rationale is given in Figure 1(b). Figure 1(b) Ameba representation for an integrated assessment of farming system Source: From Masera and López-Ridaura (2000. one can also represent the potential effects of alternative policies in relation to the changes that they could induce. where 0 refers to the worst imaginable situation and 100 stands for a ideal situation of a pristine ecosystem completely undisturbed by humans. This graphical representation makes possible to compare the existing situation characterised on a set of indicators in relation to a target situation. this would require the knowledge of: . respectively. 1992.79). The values taken by the various indicators are represented over axes moving away from the centre. In this way. For ecological applications.82) stated that “AMOEBA can serve as an adequate indicator for sustainable development by definition”. This claim is not so obvious.310). In this way. “Since water authorities and policy-makers require a clear and simple presentation a ‘radar diagram’ has been used” (ten Brink. (2000). (2002). lie inside or outside the reference circumference. In this way. For farming system analysis. In this practical example. it is easy to convey the information. then. the radar diagram represents the comparison between the current ecosystem state (reflecting a given selection of indicators) and the ‘natural state’ of the same typology of ecosystem used as a reference benchmark. see Amiba representation by Masera and López-Ridaura. The numerical values assumed to be the target for each of the various indicators are normalised. In the application illustrated in Figure 1(a). p. modified In this example. In fact.Graphic tools for data representation in integrated analysis 273 The AMOEBA is based on a radar graph. p.

4. and the expected effects of the interaction of the economic process with the ecological process. non-linear behaviour. which can imply the second type of loss. The handling of different sets of data using the AMOEBA approach does not avoid the ‘aggregation syndrome’ (a point that has been recognised by the author himself). Its goal is that of helping users to clarify their understanding of the conditions of the people living in that socioeconomic system. In the first case.g. the barometer of sustainability is a tool for measuring and communicating the degree of well-being and progress towards sustainability of a given society. Figure 2 Barometer of sustainability Source: Prescot-Allen (1996) .274 • • • T. Gomiero and M. and threshold values associated with possible catastrophic events. it is the procedure adopted for the making of the graph (e. Giampietro what is the right set of indicators to be used for such an integrated evaluation what is the right set of values to be used as a reference benchmark how to interpret the distance between the actual state and the target state in terms of performance. whenever hundreds of data gathered in non-equivalent descriptive domains have to be handled in a single graph. divided into a number of quality zones. 2001). a number of indicators of human and ecosystem well-being are aggregated into two indices and then represented in a Cartesian axes. The barometer of sustainability provides a systematic way of organising and combining indicators (Figure 2).2 Sustainability barometer According to the author Prescot-Allen (1996. genuine ignorance. In the second case. These three conditions would imply the ability to deal successfully with uncertainty. Put in another way. it is the quality of the problem structuring that affects the first type of loss. there is always the risk of: • • losing valuable information reducing the transparency and reliability of data. the conditions of the ecosystem.. the set of indicators used in this graph include only ecological indicators. Moreover. In this approach. how to normalise the values). Humans are not included in such an analysis.

Again. variables. It should be noted that due to the existence of non-linearity in the mechanisms determining the ‘health’ of ecosystems and the well-being of people living in socio-economic systems. in reality. population. Recently. • The axes are divided into qualitative equal segments representing bad. such an operation of scaling and linearisation is particularly delicate and dangerous. Comments Such an approach provides a very effective communicative tool but at the same time presents also a few relevant problems. water quality. Prescott-Allen (2001) combines in the barometer: • • 36 indicators for the Human Well-Being Index (e. When aggregating a multitude of indicators into only two indices. the value taken by an aggregate index (based on the values taken by a set of adequate indicators) indicating the degree of ecological well-being of the ecosystem in which the society is operating the axis of abscissa. global atmosphere. As noted earlier. Another interesting issue is how to know whether or not the same set of indicators can be used to compare different typologies of societal systems (whether. communication.g. we completely lose track of the information that each one of them was carrying into the representation. are derived by using the information coming from different indicators. for instance.Graphic tools for data representation in integrated analysis The Cartesian plan is built by putting on: • 275 the axis of ordinate. crime. energy use.. The two indices are then combined into a Well-Being/Stress Index that measures the degree of human well-being each country (or human-environmental system) obtains for the amount of stress it places on the environment. which can exhibit different forms of non-linearity. freedom. many indicators can be redundant (since very often indicators of development are strongly correlated). health. we cannot have a clue about what is going on both in the human and environmental systems in terms of relevant characteristics. education. fraction of protected areas. land health. peace. This is especially relevant when considering that the aggregate indices on the two axes. By looking at the final diagram. well-being in Islamic countries is perceived in a different form from that in Western countries) and . value judgement (and therefore a certain degree of arbitrariness) is involved also in the very act of choosing this set of indicators. the value taken by an aggregate index (based on the value taken by a set of adequate indicators) indicating the degree of human well being of the society. species diversity. also the clear definition of targets and thresholds on the quality zone (referring to indices) carries very limited information on what is going on in relation to the characteristics of the systems described by each indicator.. any aggregation process requires value judgements when assigning the relative weights to different indicators included in the two sets (36 and 51 indicators). poor. and mechanisms that are considered important for the analysis. and good performances. and equity) 51 indicators for the Ecosystem Well-Being Index (e. These two indices are scaled in a way to make the differences on the axes in relation to the qualitative segments comparable. water supply. according to the two indices of exploitation pressure and well-being. Moreover. OK.g. air quality. medium. wealth. As a consequence of this fact. and resource pressures).

stress in tropical ecosystems). 2001.. For this task. 2000. Gomiero and M. Four parameters are represented over the four axes of the kite diagram: revenues jobs for the human well-being spawning biomass nurseries for the ecological well-being. Giampietro different typologies of ecosystems (stress in semi-desert ecosystems vs. Garcia et al.3 Sustainability reference systems (SRSs): kite diagram Starting from the rationale of the sustainable barometer developed by Prescot-Allen (1996). indicating the position of a fishery (black polygon) in relation to four criteria (after FAO. Put in another way. . 4. modified) Two domains are considered in this graphical representation: • • • • • • human well-being ecological well-being. 2001. the incredible ability of this method to organise the information in a pattern very easy to communicate (throughout an enormous simplification in the final image) is at the same time its weakest point. he proposes a kite diagram illustrated in Figure 3 (see also Garcia and Staples. Gracia et al. FAO. Figure 3 Example of SRSs-kite diagram (isometric kite)..276 T. 1999. 1999). This is especially important when the approach is used to compare societies and ecosystems quite different in their typologies. Garcia (1997) and FAO (1997)4 has developed a Sustainable Reference System to be applied to the sustainable management of fishery.

1996. ecological..Graphic tools for data representation in integrated analysis 277 The values on each axis are normalised from 0 to 1 (in the case of isometric representation). they refer at the same time to both a quantitative and qualitative issue. As they stand. and therefore they consider experts’ opinions as sufficient for dealing with resources management in a ‘value-free’ way. the problem of aggregation remains unsolved or better untreated in clear terms. and Spash and Clayton (1997) adopting insights from system theory propose Sustainability Assessment Maps (SAMs) (Figure 4) as a tool to understand behavioural patterns of system performance. 1996. It can be used to develop a method for representing the pressure of system exploitation. social. According to the authors (Clayton and Radcliffe. As Garcia (1997) and FAO (1997. The grey scale refers to the assessment based on categories used in the representation: • • ‘good’ (clearer belts) ‘bad’ (darker belts). one-dimensional mapping. SRS is thought as a device that helps to ‘represent’ sustainability. A comparison of different system performances can be obtained by the analysis of their different graphical representations. That is. Although Prescott-Allen is clearly concerned with the value judgement that the qualitative assessment implies (even though less attention is given to the value judgement involved in the aggregation procedure and choice of indicators). This can be explained by the fact that they are concerned only with fishery (a more defined field for structuring the sustainability assessment). In this approach. Comments The definition of the terms ‘Isometric’ and ‘Anisometric’ proposed by the authors are confusing. economic). the representation: “… can be used to compare the profile of different (fishery) systems including the ‘ideal’ one with optimal values for all parameters”. Spash and Clayton.4 Sustainability assessment map (SAM) Clayton and Radcliffe (1996). SAM consists of a diagram in which each critical dimension of a complex problem is represented by an axis on a radar diagram. in fact. how figures are scaled along the axes. the effects of ‘sustainability dialectics’ (facing incommensurable trade-offs when trying to pull a too short blanket in different directions) can be made more explicit.146)4 states. different policy options are analysed according to different criteria belonging to different dimensions (e. p. Here. the authors of the SRS seem not to consider such an issue. Spash and Clayton. . Measurement of changes or indications of priorities are then mapped onto these axes. By this approach. and it can be easier for the decision makers to understand complex problems (Clayton and Radcliffe. and how they are qualitatively evaluated (by some given observers).g. more than ‘measure’ it. 1997). this approach offers a framework in which information from different descriptive domains can be integrated without being forced into a single. 1997). 4.

Tufte (1983. Zar. the figure is not easy to understand because of too much visual complexity. as well as global scale. Figure 4 Sustainability assessment map representing indicators referring to different energy policies Source: From Spash and Clayton (1997) 4. In fact. so that their visual assessment is compromised. belonging to different domains. Moreover. p. Representing together different sort of indicators. Such a selection aims at characterising the performance of the system in terms of cost effectiveness. The actual performance of the farming system under analysis is compared against the set of values reflecting desirable results.278 T. basic income/profit). it can induce visual illusions of the sort described for pie charts based on the use of sectors. regional. without a clear distinction of these domains. 1984). 1997. 1999) proposes a system of reference based on a radar diagram as a tool helping the process of defining prototypes in integrated farming system analysis. . p. nature/landscape. Gomiero and M. 1985.5 A methodical way of prototyping integrated and ecological arable farming system (IEAFS) Vereijken (1992. who claims that pie charts are: “… completely useless …” for this purpose.178). since the representation of the values taken by the indicators is affected by the compression effort. these objectives are transformed into a given set of multi-objective parameters.178) states that pie chart representation “… pie charts should never be used”. However. After establishing a hierarchy of objectives (food supply. Giampietro Comments The graphical representation is intended to represent multiple indicators over local. is another additional reason of potential confusion for the reader. Even more straight is a statement as quoted in Tufte (1983. this is a procedure for graphical representation of data that is not recommended (Tufte. Such a comparison is graphically represented by a radar diagram as in Figure 5. Cleveland. 1983.

g. 2000). On the contrary. setting the target value at 0 (no exposure). Comments This is an interesting approach to monitor farming system changes. The selection of indicators.6 Intervention impact assessment (sustainability assessment) A graphical representation based on a matrix of boxes of different colours is proposed by Efdé (1996) for a multi-criteria assessment of seven typologies of livestock management interventions (e. 1997.. as it is quite simple to understand and effective to communicate the trends in the farming system. the potassium available reserve in the soil (KAR) got worse from 1992 to 1996 resulting in a movement on the graph away from the reference point. in relation to desirable results (where the difference between the value expected and achieved = relative shortfall). it is to say the previously established target of desirable values to be achieved by the defined variables. Efdé. the diagram shows a great improvement in time (from only a few species in 1992 to 42 species in 1996)..g. Each intervention is assessed according to four different indicators: . The diagram then conveys information about: • • the relative shortfall of achieved results. economic criteria). Let us have a few examples to explain better the diagram.Graphic tools for data representation in integrated analysis Figure 5 279 Radar diagram for a prototype of integrated ecological arable farming system (IEAFS) Source: From Vereijken (1997) The external circumference of the diagram is determined by the union of reference points. 1996. Let us consider the desirable value of the parameter ‘Exposure of the environment to pesticides’ (EEP). Masera and López-Ridaura. focuses specifically on agriculture technical performance missing many other important aspects that concur to shape the structure of the farming system (e.. Udo et al. 4. however. This is described by the parameter getting closer to the reference point. In the same way in relation to Plant Species Diversity (PSD).

Example of quality matrix: intervention impact assessment Figure 6(a) Source: After Efdé (1996) modified Seven quality classes are defined by the authors. Four types of quadrants are then used to represent the achievements of the farming systems according to a previous established range of values. a number of typologies of farming systems are listed (e. On the top. On the left of the matrix diagram. and the quadrants of the matrix (the boxes of different colours) are coloured according to the categories given in the legend. As acknowledged by some authors Udo et al. integrating cereals with orchards and vegetable production). cereal production. (1997). Gomiero and M. Nevertheless.. Giampietro total production productivity of labour productivity of capital environmental impact on soil (Figure 6(a)). it can represent a useful tool to structure a participatory discussion on the sustainability of a farming system.g. Masera and López-Ridaura (2000) use a mix technique for the presenting integrated results (Figure 6(b)).280 • • • • T. a number of relevant indicators are listed to characterise the farming system typology. Their sustainability assessment diagram is a way to combine graphic representation and numerical data. . integrating corn and beans. the Intervention Impact Assessment approach has to be considered as a rather approximate qualitative tool. This approach may be useful to provide rapid and easily comprehensible data to farmers.

g. Ecosystem Quality 40%. the point is positioned where Human Health is weighted 50%. 2000 – based on Hofstetter. Each point within the triangle represents a combination of weights that add up to a 100%. (1999) for Life Cycle Assessment (LCA) analysis (Figures 7(a) and 7(b)). The representation also is quite rough and does not consent to represent actual values of system performance.g. and Energy Resources 10%. A key feature is the possibility to draw lines of indifference. triangular graphics of this sort have been in use for many decades in other disciplines (e. In the example of Figure 7(a). comprehension can be somehow compromised. mineralogy. These are lines representing weighting factors for which product A and B have the same environmental loads.Graphic tools for data representation in integrated analysis Figure 6(b) 281 Diagram for the presentation and integration of results of a sustainability evaluation Source: From Masera and López-Ridaura (2000) Comments Graphical representation based on a matrix of boxes results easy to comprehend. soil science.. The triangle can be used to graphically depict the outcome of product comparisons for all possible weighting sets. 1998). .. The lines of indifference divides the triangle into areas of weighting sets for which product A is favourable to product B and vice versa. However. 4.7 Mixing triangle The use of Mixing Triangle graphical representation has been proposed by Hofstetter (1998) and Hofstetter et al. material science). geology. anyway when the number of levels considered increases (e. It can be anyway useful to easily manage qualitative information or rough quantitative figures. The position of such a point is defined by following each side until the dotted flashes leave towards the point in the triangle (Pre-Consultants. Figure 6(b)). in particular in contexts where stakeholders can have problems in the comprehension of more complicated graphical representations.

p. That is. in the final assessment. 2001. (1999). More information on this subject can be found in Hofstetter et al. Also in this case.1). . This methodology can facilitate an open discussion with the stakeholders. instead of a tool that produces simple single statements. this representation is very useful. In spite of this fact. Comments Dewulf and van Langenhove (2001) point out that such methodology presents some important problems. In fact. but they have to agree whether it is plausible that the weights would fulfil the conditions under which A is better than B or not. Such an approach therefore turns LCA into a consensus building process. Both the set of units and the scale can differ from one item to the other. The line of indifference in the weighting triangle and the sub-areas with their specific ranking orders (B > A means that alternative B is environmentally superior to A and the eco-index A is higher than B). we face the same problem found with the approach proposed by Prescott-Allen.282 Figure 7 T. (1999) According to Hofstetter (1998) and Hofstetter et al. in the aggregation process we ‘get the point but lose the system’. (1999). The various effects related to the life cycle of a given product are measured on a particular scale with a given set of units. Gomiero and M. since it enhances the transparency of the weighting process. these numerical measures have to be reduced (assuming full comparability) in order to obtain a unique final assessment as result. “The balancing process is … a rather subjective and arbitrary step in LCA methodology” (Dewulf and van Langenhove. The stakeholders do not have to set discrete weights. it shows under which conditions (which weighting factors) product A is better than product B. They argue that. Giampietro Mixing triangles for integrated (multiple criteria) representation of environmental systems Source: From Hofstetter et al.

Spread category conveys an indication on the inexactness of the information in the numerical and unit places (statistical notation). Pedigree conveys an evaluative account of the production process of the quantitative information..php?op=viewarticle&artid=3 Pedigree conveys an evaluative account of the production process of information and indicates different aspects of the underpinning of the numbers and scientific status of knowledge used (Funtowicz and Ravetz. Assessment should express a judgement of the (un)reliability associated with the quantitative information conveyed in the previous categories. etc. mentioned previously. or alternatively through those of Bayesian statistics. 2000. It maps the state-of-the-art of the field in which the quantity is produced. or a set of elements and relations expressing magnitude (e.). This category operationalises the epistemological sort of uncertainty.net/sections.g.net at http://www.Graphic tools for data representation in integrated analysis 283 4. fraction. Risbey et al. large. borders with ignorance.92 for the . Figure 8 Kite diagram maker for NUSAP Source: Available on the website nusap. or GNP per capita). p.. van der Sluijs et al. Unit represents the base of the underlying operations expressed in the numeral category (can be divided into standard and a multiplier) (grams.. 2001. 1990.8 Kite diagram for NUSAP applications NUSAP is a novel approach to uncertainty assessment proposed by Funtowicz and Ravetz (1990).. It may be represented through ‘confidence limits’ and ‘significance level’ of classical statistics.nusap. 2002. or ordinal indexes sometimes expressed in verbal locutions – small. van der Sluijs et al.. Kite diagrams (Figure 8) have been used within NUSAP approach for pedigree and uncertainty assessment of data on SO2 and CO2 emission and scenarios (Corral Quintana et al. in which: Numeral entry may be a number. 2002). intervals. decimal digits. NUSAP is an acronym for the 5 categories: • • • • • numeral unit spread assessment pedigree.

This is again an example of how difficult is to forced compression of different incommensurable criteria. (2002).. Gomiero and M. Figure 9 Pie representation for the policy performance index (PPI).9 Pie for policy performance index (PPI) Anther sort of graphical representation is that used by Jesinghaus (1999) at the EU Joint Research Centre at Ispra (Italy). 4. so that if used alone to represent the quality of the process. It should be noted. method. proxy. In the case of van der Sluijs et al. However. the criteria used are: Proxy (it refers to how good or close a measure of the quantity that we model is to the actual quantity we represent). Corral Quintana (2001) uses radar diagram to visual representing Pedigree of data (concerned with the quality of the used information. Methodological rigour (it refers to the degree of reliability of a specific methodology). after Jesinghaus (1999) . in fact. and the influence of the decision tools). Giampietro kite diagram). As pedigree assessment involves qualitative expert judgement. An observation should be made concerning the creation of an overall qualitative numerical pedigree averaging out the linguistic description of the four-level. 2001) for the same assessment (CO2 emission). discrete. numerical scale (0: week. the role of the analyst.284 T. Within NUSAP approach. The representation aims at allowing the citizens be able to judge ‘at glance’ the government’s performance on a broad range of issues. Using kite diagram representation in this case is appropriate as it supplies an easy understanding of the evaluation exercise. we use only four of these: validation. An example of the approach is provided in Figure 9 (different types of this representation are used from the author). numerical scale (0: week. NUSAP based approach. Comments This approach is very interesting for it forces the researcher to address issues that are usually missed or overlooked and that concern the quality of the process of data construction. Theoretical understanding (it refers on how well are theory established in the scientific field). Validation (it refers to the degree to which it has been possible to cross-check the data and assumption used to produce the numeral of the parameter against independent sources). to 4: strong) on the qualitative pedigree of information. and empirical basis. to 4: strong). discrete. it loses significance. that the same comprehensive index can result from a number of value combinations. Empirical basis (it refers to the degree to which observation and statistics are used to estimate the parameter). for instance (Risbey et al. a different combination of criteria can be used. uses a linguistic description on a four-level.

285 In the detailed representation. economic indicators (40%) count more than environmental ones (25%). A similar index is also created to cover social issues.e. inflation rate. climate change.. The size of each segment. Whenever this representation is a tool used to inform citizens about government performance in solving the perceived problems. it is well evident that the choice of indicators is a very delicate matter. each of which is composed of six indicators (i. in this representation. GDP is an index composed of several hundred indicators weighted by market prices). the environmental pressure index is sub-divided into ten ‘policy field indices’. and another index is made for the economic performance and consists of typical indicators such as GDP. a total of 60 components).. etc. The inner two levels are aggregated valuations of the underlying segments (i.e. education. reflects the importance (the ‘weight’) of the issue for politics. The colour of each segment reflects the judgement of performance using a seven-colour scale. income distribution and poverty. the government should focus on improving the one with a higher influence on the colour of the ‘Environment’ segment. like the quality of health services.Graphic tools for data representation in integrated analysis The pie is divided into three sectors representing: • • • economy environment social care (assuming that the relative indicators are available and widely used by the media). i. All three indices are then aggregated to a Policy Performance Index (PPI) and presented as a pie chart organised in three concentric circles as follows: • • • one overall index (PPI) in the centre of the pie three sub-indices for economy. Comments Although at first sight the mechanism looks a bit complicated. and environment an outer circle representing sub-sub-indices or ‘simple’ indicators such as: GDP (even though technically speaking. and air pollution pressure index. inflation.e. and investment rates.. green for ‘good’ and red for ‘bad’. then. it becomes a communication rather than a working tool. . it provides the government with two simple rules for their decision making: • • you must eliminate the red spots: voters do not trust governments that are unable to solve a crisis or to deal with a very bad situation indicators with high weights have a high political priority: in the cases of two environmental signals of ‘crisis’. social care.. the ‘yellow’ = ‘medium’ PPI shows the average of the underlying valuations ‘good + bad + very bad’). waste. The construction of the pie is made through polls asking people about the most urgent problems and how the government is solving them. poverty rate. In this example.

. Vreeker et al. 1998. The impact matrix is formed by the values that the indicators (from economic. Gomiero and M. Besides the construction of the impact matrix. These values represent the reference system for judging the alternatives. There are three approaches to such a representation: • • • qualitative quantitative hybrid approach. These values represent the reference system for judging alternatives. which are complementary one another. preferably a critical threshold value (CTV) has to be defined. Nijkamp and Vreeker. As already discussed in the Section 4. 2001). The flag model uses two types of input: an impact matrix and a set of critical threshold values. The flag model has been designed to assess the degree to which competing alternatives fulfil pre-defined standards or normative statements in an evaluation process. Zar. and environmental domains) assume for each alternative considered. a bandwidth of critical threshold values is constructed. 1984) strongly advise not to use pie charts to convey complex information. 1983. Cleveland. . This can be represented as follows: Section A Section B Section C Section D Green Yellow Red Black No reason for specific concern296 Be very alert Reverse trends Stop further growth The flag model can operate both as a classification procedure and as a visualisation method. social. 4. For each indicator in the flag model. 1985. scholars in the field of data representation (Tufte.3. There are four important steps in applying the model: • • • • identifying a set of measurable indicators assessing the impact of the alternatives on the above-mentioned indicators establishing a set of normative reference values (standards) evaluation of the relevant alternatives.286 T. Giampietro The problem with this representation is that it is difficult to order the values encoded on the pie chart from smallest to largest. Since in many cases experts and decision makers may have conflicting views on the precise level of the acceptable threshold values.10 The flag model The flag model has been develop by Nijkamp and colleagues with the purpose to analyse whether one or more policy alternatives can be classified as acceptable or not in the light of an a priori set of constraints (Nijkamp and Ouwersloot. for each indicator a critical threshold value has to be defined. Relative sector dimensions are then difficult to be adequately perceived.4 for the sustainability assessment map. 2000. This bandwidth ranges from a maximum value (CTVmax) to a minimum value (CTVmin).

That is to say. county or local. this can greatly affect the overall results. Gomiero. Gomiero and Giampietro. Giampietro. The quantitative approach defines the values of the standards that may be acceptable or not (for further details. the normalisation process and the setting of critical threshold value requires as much transparency and detailed explanation.. 2001. social. Pastore et al.. has been proposed by Giampietro and colleagues (Giampietro and Pastore. .g. if many criteria are selected in the economic domain and few in the environmental (or vice versa). by adopting the ‘flag model’ proposed by Nijikamp and colleagues.11 Multi-objective integrated representation (MOIR) A particular organisation of data in a graphical representation. but it has been developed as to include two key steps: • the message of normalisation of values for communication of information (by dividing the area of the radar graph in zones indicating bad. Likewise for any other methods of this kind. In fact. Vreeker et al. Nijkamp and Ouwersloot. However. Comments The flag model offers a very interesting potential for representing integrated assessment of environmental systems in a way that can be helpful for the stakeholders. and. global label). 1999. as the number of the flags gained by a domain (economic. even more importantly • the benchmarking process over the set of indicators. 2000. This graphic representation was developed starting from the basic idea of the AMOEBA (using different criteria and indicators for different scales). 2003. the viability domain (or other sets of benchmark points). 4. 2004. has to be established in the sense of direction. Nijkamp and Ouwersloot. which of the extremes have to be regarded as preferable to represent the best performance according to some specific goals. this still leaves much to subjectivity of the assessment. see for instance Vreeker et al. Nijkamp and Vreeker. Once defined.g. household.g. 1999. 1998. This implies the explicit introduction of a value judgement that can be given both by the analysts (outsiders the system) and stakeholders (insiders the system).. 1998. 2004) (Figure 10). The qualitative approach only takes into account the colours of the flags and merely displays in various insightful ways the results obtained from the evaluation (this approach is adopted in the MOIR representation illustrated in the following Figure 10). for a given indicator of system performance. 2001. medium and good performance – e.Graphic tools for data representation in integrated analysis 287 This allows for the method to be flexible to the requirements of its users. called multi-objective integrated representation (MOIR). regional. village. (e. MOIR has the goal of establishing bridges across non-equivalent representations referring to patterns perceived and detected on different hierarchical levels (e. Nijkamp and Vreeker... This because it faces the challenge posed by the acknowledgement of the existence of trade-offs in decision making.. 2000). 2001. 2001). environmental) depends on the number of criteria considered for each domain.

Of course. Giampietro Figure 10 MOIR application: a multiple criteria comparison of Italian and Chinese freshwater aquaculture. changing the sets of indicators the patter will be modified. Anyway. human time. which can be more or less characteristic for a given farming system. energy. Comments MOIR can be a useful tool to: • • force the analysts. . that is to say. by using congruence over flows of money. in order to be able to discuss of the trade-off associated to different scenarios. as well as if we choose a scale instead of another and so on. Integrated representation means: • • • establishing links among processes occurring on various contexts and levels (e. to see the problem under multiple dimension to recognise pattern of functioning in the system under analysis.g. after Gomiero and Giampietro (2003) By the integrated representation. as well as the stakeholders.. it is possible to establish links among the value taken by different indicators. A system will be much better characterised by some given patterns.288 T. This holds true also for other systems. it has to be recognised that the option space and the constraints posed to a given system from its lower and higher hierarchical levels do not allow whatever system structure to take place. and matter) establishing links among relevant criteria focusing on the existence of trade-offs. to think in a systemic way. Gomiero and M.

By using the same rationale. it provides information not only on the values of the individual variables included but also a sort of ‘gestalt’ view of the whole (Funtowicz and Ravetz. colours. This (Funtowicz and Ravetz. Chambers. Giampietro and Pastore. 5. 1990. 1980). • In fact. 1984). Giampietro et al. 1992. e. 2004. Giampietro. scales. it is possible to establish a link between changes described in an integrated (considering multiple criteria) graphical representation and changes in land use. Epistemological issues related to the very same process of figure and diagrams making. Gestalt psychology (Köhler. Gallopin. Vereijken. Corral Quintana. the very same act of selecting a scale to represent a set of data. when dealing with sustainability dialectics). 1990. etc. 1996). 1997.Graphic tools for data representation in integrated analysis 289 5 • A critical appraisal of these graphical representations Graphical representations can have problems because of: The psychology of human vision and its organisation and interpretation as performed by the brain. Generate a dynamic graphical representation of changes in indicators when discussing scenarios By establishing relations between changes in biophysical variables and economic variables. it can be possible to describe with models the possible effects – or better the feasibility domains – in terms of values taken by a given indicator in relation to another (Giampietro. Kanizsa. Asch. 1999. 1997). 2001. which do not have compensation. 2004) Make detectable some properties of the whole not easy to detect for non-experts (Gestalt) Graphical representations can provide a profile of system performance (within the chosen set of criteria and indicators). can determine in the reader different feelings and generate different feed-backs (Zar. Nijkamp and Ouwersloot. 1998.. The crucial role that the graphical representation can play in a process of multi-criteria evaluation points at the consequent problem of individuating ‘quality’ criteria for making them (Funtowicz and Ravetz. 1997a. This aspect is particularly important in participatory processes where information has to be understood by a wide range of people.. 1990. Gallopin. 1996). Software for a . 1997. 1992). pp.g. we can: Convey relevant information in a form easily comprehensible to the stakeholders Graphical representations make possible to have a clear and simple presentation of technical data often required by policy makers and stakeholders in general (ten Brink. 1947. than can be thought as a sort of ‘map’ of the system performance in relation to different set of indicators referring to criteria not-optimisable all at the same time (e. Distorted interpretation of graphical images is a field much studied in psychology. 1997b. use of particular forms.g. 1968.83–98) provides several insights about the practical usefulness of maps and diagrams in delivering information for policy. 1979. and also in relation to the generation of possible feedback (Nijkamp.1 Pros of graphical integrated representations By mean of graphical representation. e.. As such. 2001.. 1997. Spash and Clayton. This can be used to have a look at the ‘winners and the losers’ so to speak in the final negotiation over those sustainability trade-offs.g.

However.2 Cons of graphical integrated representations Graphical representations are not to be expected. On the contrary. Gomiero and M. there is the .. • Those aggregation procedures aiming at getting a final index of performance in the step of representation do not have the ability of keeping these two steps separated. These consequences are expressed according to the set of indicators that better reflect the concern of the various social groups involved in the process (admitting that a participatory process was employed to select the identity of the multi-criteria space). in spite of the caution due to the short history of this field. In this way. It should be noted that this graphic representation is very flexible for handling any type of analysis requiring the simultaneous consideration of different criteria. it is possible to keep separated: • the phase of the understanding of the implications of different options (according to both the perceptions provided by the scientists and those provided by the stakeholders) the phase of decision making in which the indications given by various indicators have to be mediated in relation to the effects expected on each social group (direct relevance) and to the effects that they will imply on other social groups (indirect relevance). Especially problematic is the case of integrated graphical representations having the goal of handling indicators referring to different scales and incommensurable criteria.g. Vereijken. Corral Quintana. 2001. extreme caution has to be adopted when using these innovative tools.. Obviously. to be ‘intuitive’ or ‘self-explanatory’. Clark et al. they present a wide array of problems both in the making and in the comprehension of the different graphs.. 2000. due to the inherent complexity of the interaction between socio-economic systems and ecological systems. 5. Giampietro. 1992. Gomiero. they make explicit the consequences (both in positive and negative) implied by an alternative (or differences found in systems compared in the analysis). 2001. Obviously. 2004. a dynamic integrated assessment of scenarios seems to represent a very promising direction of development. Put in another way. de Koning et al. Hall et al. in general. 1992.. 1997. a ‘tailored’ representation of the profile of distribution of ‘costs’ and ‘benefits’ over the social groups (represented considering both the perceptions provided by the scientists and those provided by the stakeholders) can represent a valid tool to facilitate an open discussion over contrasting views about the set of alternatives (ten Brink.290 T. In all those cases in which the graphic representation is used to aggregate the various indicators into a single index (this requires that weighting factors be applied to the various indicators). Giampietro and Pastore. In this way. Spash and Clayton. Giampietro dynamic graphical representation of land use change does exist and can be used to discuss of scenarios (e. this applies only to those representations that keep the set of various indicators (which are relevant and significant to a different degree for different social groups) in the final graph (rather than collapsing them into a single overall index). 1995. 2004). 1998. Facilitate the discussion on incommensurable trade-offs (effects of sustainability dialectics) Graphical representations help the stakeholders in visualising the implications of sustainability dialectics. Verburg. 2000).

despite all their problems. which implies a necessary cost. Obviously.g. 1998. In order to be able to make a decision in a finite time. criticism on methods of graphical representations has to be based on the consideration of the trade-off between ‘loss of relevant information’ and ‘gain of usefulness in the organisation of the information space used for problem structuring’ that the various procedures entail. the problem is certainly a serious one. The problem has been already pointed out for graphical representations in the field of Life Cycle Assessment (Hofstetter. The choice of a given set of relevant criteria. This ability to involve the stakeholders is crucial.. different choices of identities for the elements to be adopted in a model. 2000). and on them of targets and admissible ranges.. graphical integrated representations support multi-criteria participatory processes of evaluation and should be preferred to more conventional approaches such as chrematistic Cost-Benefit Analysis. In fact. of a given set of indicators. but it is common to all types of representation (and all types of models). As already noted when commenting on individual methods. it is crucial to know how relevant for decision making was the information lost in the process. Can be used to mislead the perception of a given situation Also in this case. since in a dynamic reality the perception and representation of both: • • problems expected consequences of solutions are continuously changing in time with strong non-linearity and unpredictable twists. can imply a structuring of the problem that does not necessarily reflect the perceptions of the various stakeholders. In this view. the information space used for the problem structuring has to be reduced as much as possible.Graphic tools for data representation in integrated analysis 291 obvious risk of losing track of the information carried by the original indicators. In this case. As discussed in the first two sections. The goal is to try to avoid losing too much relevant information. In relation to this challenge. This goal should call for an attempt to preserve as much as possible the original information available from the various set of data used as input. implies the loss of a lot of useful information in the final graphical output. Simplification and compression in the demand of information used in a process of decision making is a necessary step. we are discussing here graphical representations acknowledging from the beginning the integrated nature of the analysis (e. Can lead to an oversimplification of the reality This problem is certainly true. it is our opinion that these conventional systems lead to even more simplified results (since all relevant variables are collapsed in a single index with dubious procedures of aggregation). Therefore. acknowledging the necessity of structuring the representation on multiple criteria or multi-objectives and multiple scales). and on the other hand. another basic goal of graphical representations is that of involving stakeholders in a ‘quality check’ of themselves (to check the discrepancy between the representation provided by the scientists and that agreed-upon by the social actors). Hofstetter et al. the aggregation of indicators on one side helps a holistic vision in the characterisation of the problem structuring. different choices of . 1999. which moreover are not open to a quality check from local actors. Pré Consultants.

Establishment of links among different indicators. ten Brink. 1994. which have to be considered in parallel to have a meaningful analysis the individuation of different levels of analysis required to cover relevant information the various criteria that should be considered within the various dimensions. 2004. Giampietro and Pastore. dimensions. Assessment of the values taken by those indicators in relation to the alternatives considered in the analysis. 2004). 1981. Definition of a set of indicators that can represent the performance of the investigated system in relation to the set of objectives. and then of encoding variables lead to different representations of system’s profile. Also in response to this objection. according to relevant stakeholders. 1999. and levels considered as relevant in step 1. 1997.1 The steps to be followed for a sound integrated representation In conclusion. 1992. Gomiero. in order to be able to: • generate mosaic effects to increase the robustness and reliability of the information space • make explicit the implications of ‘sustainability dialectics’ associated to different scenarios. more or less. This is the reason why we envision the use of graphical integrated representation in a process of decision making.. Gomiero and M. Munda et al. criteria. FAO. 1997. Munda. we can only observe that this type of problem is common to any form of representation and problem structuring of a real situation. 2001. 1999. 2004. 1997. 3 4 The definition of such a procedure and related steps is. We wish also to warn the reader that these graphical representations cannot be used as overall assessments of the system performance. This step obviously must reflect the peculiarity of the local context and the availability of data gathered in the study. Giampietro.292 T. only within a participative procedure. 6.. Garcia. 6 Two key issues for a sound integrated graphical representation We wish to end this overview by presenting what we believe can be a sound procedure for generating integrated graphical representation. López-Ridaura et al. Giampietro observable qualities. . This requires first of all: • • • • • 2 an institutional analysis to study the set of relevant actors affected and affecting the decision to be taken the definition of the objectives related to the process of decision making the definition of the dimensions of the sustainability predicament.. the procedure for generating a sound integrated representation is based on: 1 Definition of the relevant objectives that should be considered in the integrated analysis. 2002. present in the work of different authors proposing the adoption of integrated representation in different fields (ten Brink et al.

. and looking at the consequent graph. we must apply to the various indicators weighting factors in relation to the specific situation considered reflecting the preferences. and therefore has nothing to do with an objective assessment of the overall quality for the system! The profile of weighting factors used to compare the indications provided by the set of indicators used for the integrated representation is location and time specific.g. fears of the stakeholders. 1987. However. 1990. That is. we can start the process by adopting as a first input an ‘etic’ perspective (from the system outsiders) and only after having achieved an ‘emic’ perspective of the problem structuring – agreed among the stakeholders (the system insiders) – it is possible to get into the step of negotiating weighting factors (‘etic’ and ‘emic’ are terms taken from anthropology that well address the issue of world representation asking ‘from which point of view?’ – (Harris. one could be led to believe that given the normalisation over the values taken by the indicators over the various axis and given a common direction of performance on the various axis (e. For example. 1997). In fact. In order to compare the indication given by a set of non-equivalent indicators (that can be both quantitative and qualitative) referring to different criteria. At this point. the total area included inside the profile of performance should be considered as an index of overall quality for the system.Graphic tools for data representation in integrated analysis 293 6. Chambers. For example. since different social groups (or different systems of knowledge) can express different profiles of weighting factors for the same set of criteria. in a general discussion about how to characterise and assess on a multi-criteria problem structuring the environment in which one wishes to operate. Headland et al. one can assign a crucial importance (very high weight) to a healthy air quality. • . aspirations. an integrated graphical representation provides a quality profile in relation to a specific set of indicators and criteria considered by the analysts as relevant for determining the performance of the system. This is not correct for several reasons: • The various indicators refer to non-commensurable criteria and therefore the process of normalisation does not imply that they have been weighted in relation to their relative importance in determining the overall performance of the system. when adopting a radar diagram. the final profile of weighting factors that will be adopted in the decision is the result of a negotiation (power relation) among the different perspectives. the very same person can then decide (a few minutes after having expressed such a preference) to enter into a building in fire filled with dangerous fumes to save children trapped inside.. This profile of weighting factors can only be obtained after a discussion with stakeholders based on a first tentative integrated representation (a tentative input to start the iterative process).2 Integrated graphical representation cannot be used as an overall assessments This is an objection that deserves the maximum attention. the more distant from the centre the better).

the aggregation of indicators in a singe figure. many factors can concur to introduce biases in graphical representation. the final decision let to social agreement upon the conflicting interests. the selection of the relevant criteria and useful sets of indicators). In this sense. For this reason. and should be. the particular identity of any particular graph has very little chance to remain the same when going through a participatory process of integrated evaluation. As already noted. if the fact that integrated graphical representation cannot be used as an overall assessments can be considered a weak point. being very complicated mathematical algorithms. Gomiero and M. it implies the loss of a lot of useful information in the final graphical output (loss of the holistic view). these graphical representations are useful means to provide an integrated overview of the system under analysis (of course for each tool its pros and cons can be. That is the reason why it is crucial to know the process of number construction and how relevant for decision making was the information lost in the process. and their viability domains etc. . if on one side helps to get a simplified characterisation of the problem structuring. indicators. As with the ‘creation’ of numbers. and on the other hand. This can facilitate social discussion among different stakeholders.. They are characterised by: • • • considering multiple criteria and multiple indicators within them limiting themselves to the representation of the values taken by the indicators within some viability or feasibility domain aiming at presenting patterns of the system behaviour in relation to the selected sets of indicators so to easily represent the multiple trade-offs involved in the decisional process. debated). we have presented a review of integrated (multiple criteria) graphical representations used in the field of farming system and environmental analysis. are at the core of any system of analysis (not talking about the power exerted by vested interests. it is important to make as clear as possible the process by which the graphics have been generated (this starts with the very same setting of the problem under question. it should be noted that subjectivity in selecting constrains. it has to be stressed that aggregation procedures.). This means that any graph providing an integrated representation of a situation should be considered as just one of the possible inputs to be adopted for an integrated (multiple criteria) problem structuring.294 • T. Put in another way. aim at getting a final index of performance (or the optimal solution) do not to escape the problem. stakeholder pressure. Getting a final figure out of a long and hard process of compressing makes the reader lose track of the heterogeneous characteristics of the systems. 7 Conclusions In this paper. so that it is easy to handle. Anyway. etc. objectives. Furthermore. Giampietro The profile of performance resulting from the integrated representation on a multi-criteria space refers to just one of the possible integrated representations of the system.

Girardin.L.175. Dordrecht. S. (1996) How Nature Works: The Science of Self-organization Criticality. INCO DEV Program (within the work package ‘developing analytical tools’ managed by INRAN). Method. Berlin. W.E. Bossel. M. S. p. (Ed. P. p.F.372. Agricultural Systems. Bland. C. Kaltschmitt. D..261–270. Altieri. 3. Beinat.N. in Sills. International Institute for Sustainable Development publisher. and Nijkamp. Chicago. Westview Press.A. Springer-Verlag. References Allen.B. Boulder. W. T.738. and Starr. pp. p. The Netherlands. M. Wipening-Manitoba.pdf. Macmillan company & The Free Press. S.660.158–175. London. p. p. in Härdtlein. Springer-Verlag. and Wurl. Belton. Bana e Costa. AgBe Publishing.69–83.212.Graphic tools for data representation in integrated analysis 295 We believe that the usefulness of sound graphical representations lies in helping social debate and rising awareness among stakeholders about critical trade-offs and the complex nature of living systems.L. Beets. as well as in forcing both researchers and stakeholders to make clear to themselves and to the others the process by which they represent their own reality. T. Earthscan. Asch.) (1990) Readings in Multiple Criteria Decision Aid. (1997) ‘Use of agro-ecological indicators for the evaluation of farming systems’. pp. V. Bockstaller. Holland.C. T. Dordrecht. Bockstaler. H. Kluwer Academic Publishers.396. I. No. (1987) Agroecology: The Science of Sustainable Agriculture. New York. (1990) Raising and Sustaining Productivity of Smallholder Farming System in the Tropics. 7. P. Berlin. 6. (Eds. pp. University of Chicago Press. 60. p. Lewandowski. The paper has been written within the activities of the research project ‘Southeast Asia in transition’.H. Bak.. (Ed.123. H.. (1982) Hierarchy – Perspectives for Ecological Complexity.. pp. Vol. and Girardin. M. (2002) Multiple Criteria Decision Analysis: An Integrated Approach. p. and Morse. . (1999) Indicators for Sustainable Development: Theory.ca/pdf/balatonreport.310. Eric Schimdt Verlag GmbH & Co. DG Research. New York. E. Bell. (Eds. and Steward. Canada.G. European Journal of Agronomy. C.H and Van Der Werf. P. Vol. (2000) ‘Agro-ecological indicators – instruments to assess sustainability in agriculture’. p.157–167. Kluwer Academic Publisher. coordinated by the Institute for Interdisciplinary Studies of Austrian Universities in Vienna (IFF). (1999) Sustainability Indicators: Measuring the Immeasurable. and supported by the European Union.187. (1968) ‘Gestalt theory’. Application. (1999) ‘Toward integrated assessment in agriculture’.) (1998) Multicriteria Analysis for Land-use Management. C. P. available on internet at URL: http://iisd. The European Union or the European Union Commission cannot be held responsible for results and opinions quoted in the text. Alkmaar. Acknowledgment We wish to thank two anonymous reviewers for their helpful comments and suggestions.) Nachhaltigkeit in der Landwirtschaft.M. H. Vol.H. p.) International Encyclopaedia of the Social Sciences.

p.M. pp. p. P.J. Editions De L’universite De Bruxelles.C. Agricultural Systems.R. (2001) Una metodologia integrada de exploracion y comprension de los procesos de elaboracion de politicas publicas. East Java. p. 17.791–799. pp.318. Veldkamp. and Guimarães Pereira.H. FAO. Agriculture. Technical Guidelines for Responsible Fisheries. A. Rome.. Bilthoven. pp.S. (1995) Evolutionary Dynamics and Sustainable Development. (1998) ‘Land use in ecuador: a statistical analysis at different aggregation levels’. Kluwer.B. S.. R. The Netherlands.be/~bmaresc/PromWeb.E. M. pp. Funtowicz. (1997) Whose Reality Counts? Putting the First Last. Perez-Trejo.O. pp. D. H. FAO (1999) ‘Indicators for sustainable development of marine fisheries’.R. A System Approach. A.323. Wageningen Agricultural University. Clayton. pp. Edwar Elgar. S. Corral Quintana. S. Rome.htm. A. Cleveland. Journal of the American Statistical Association.90.828–833. (1985) The Elements of Graphing Data. 5. p. H.230. W. London. available at URL http://student.E.147–156.pdf. and McGill. G. and Marescal. Funtowicz. de Koning. Earthscan. (1990) Uncertainty and Quality in Science for Policy. . 229. Une méthodologie d'aide à la décision en présence de critères multiples. Ecological Economics. R. ISBN 84-688-4314-8. (1996) Quantified and Integrated Crop and Livestock Production Analysis at the Farm Level: Exploring Options for Land Use of Mixed Farms on Heavy Limestone Soils South of Malang. J-P. (2001) ‘Assessment of the sustainability of technology by means of a thermodynamically based life cycle analysis’. A. pp.ac. Conway.248.jrc.680. p. Efdé. Science. de Zwart. 70. J. Giampietro Box. Dordrecht. 8. p. UK.175. (2000) Exploring the Quality of Environmental Decision-making Processes. D.231–247.pre. Beacon Press.213. B. Ecosystems and Environment. National Institute of Public Health and the Environment. p. p. Daly. p. Aldershot. L. Vincke. (1987) ‘The properties of agroecosystems’.296 T. and van Langenhove.228–238. Bruxelles. (1989) For the Common Good: Redirectioning Economy Toward Community. R. (1995) Manual on Integrated Water Quality Evaluation. available at URL: www. Boston. (1986) ‘How to select and how to rank projects: the PROMETHEE method for MCDM’.188. 24. Universidad de la Laguna. Brans. Vol. FAO (1997) ‘Land quality indicators and their use in sustainable agriculture and rural development’. 24. (1996) ‘On the concept of weak sustainability’. Vol. Vol. S. W. J-P. (1996) Sustainability – A Systems Approach.org/ag/magazine/GFSS.S. p. Environment and Science Pollution Research. The Netherlands.1–7.482. Tenerife. Clark. and Gibbon.it/pedigree/pedigree. 8. Vol. Summit. and Cobb.P. FAO. 71. S.nl/ eco-indicator99/triangle. F. The Environment. and Mareschal. B. in Spanish. and Allen.258. Brans. European Journal of Operational Research.. Vol. available in internet at URL: http://alba... Vol.J. (1985) ‘Graphical perception and graphical methods for analyzing scientific data’. Gomiero and M. p. Hobart Press. G. B. and Ravetz. Belgium. and The Sustainable Future. and Fresco. p. J.229.fao. No. Cleveland. PhD thesis. p. Dixon. Vol. J.68. J. Indonesia. and Radcliffe. Chambers.O. London. J. Vol. (1976) ‘Science and statistics’. Gulliver.vub. N. Dewulf. Cabeza Gutés. and Mareschal.html. FAO Land and Water Bulletin. 3. Brans. (2002) PROMETHEE-GAIA. NJ.P. available in internet at URL http://www.H. Ph. G.htm. N. (2001) Global Farming System Study: Challenge and Priorities to 2030. (2004) How to Decide with PROMETHEE. Vol.95–117. Corral Quintana. Intermediate Technology Publications. and Trivedi.

(1997a) ‘Socioeconomic pressure. Boca Raton. Freshwater Res.187–199.Graphic tools for data representation in integrated analysis 297 Gallopin. and Martel. Headland. Vol.F. CRC Press. in Ulgiati. and technological changes in agriculture’. p. pp. and Giampietro. Giampietro. 43. in Freese. pp. Giampietro. Environmental Modeling and Assessment.C. M. Population and Environment. New York. Bukkens. Vol. and Environment. p. pp. G. Staples. 330. Gliessmann.537–556. Giampietro. C. pp. p. London.G. G. Headland. and Pastore. PhD thesis in Environmental Science.) (2002) Panarchy: Understanding Transformations in Human and Natural Systems.H. in Gliessman. (2004) Multi-scale Integrated Analysis of Agroecosystems: A Complex System Approach. Giampietro. multiple scale performance space’. Vol. and Staples. (Ed.501–521. D. S. and Environment. S. Vol. K. p. T.244. G. Hall. Gunderson..S. C. Pike.) (2001) Agroecosystem Sustainability – Developing Practical Strategies. M. JAI Press. S.A. A summary of the book by T. p. CRC Press. is available at URL: http://www. (2001) ‘The FAO guidelines for the development and use for sustainable development of marine capture fisheries and an Australian example of their application’. Giampietro. pp.sil.201–229. Agriculture. 18. J-M. M. Ocean & Coastal Management.htm. pp. T. and Chesson. in FAO (1997). Lewis Publisher. (2000) ‘Sustainability reference systems and indicators for responsible marine capture fisheries: a review of concept and elements for a set of guidelines’.129–199.N... Critical Reviews in Plant Sciences. S. Mar. No. Vol. No.385–426. Spain. 2nd ed. Vol.437. natural resources.384. Harris.) (2000) Quantifying Sustainable Development: The Future of Tropical Economies. M. (1996) ‘Environmental and sustainability indicators and the concept of situational indicators. (1997) Indicators for Sustainable Development of Fisheries. Gliessman. and Harris.299.S. Giampietro.J.): Agroecosystem Sustainability-Developing Practical Strategies. Vol. (2001) ‘Multiple-scale integrated analysis of farming systems: the Thuong Lo commune (Vietnamese uplands) case study’.506. S. (1997b) ‘Socioeconomic constraints to farming with biodiversity’. D. 62. New York. L. (Ed.R. London. and Leon Perez.203. 65. L. and the socio-economic structure of human society: examples and applications’.177–202. Gomiero. Harper & Row. M. (1987) Cultural Anthropology. Garcia. (2000) Agroecology: Ecological Processes in Sustainable Agriculture. and Pimentel. p. T. . (Eds. D. (1998) ‘Tentative guideline to help choosing an appropriate MCDA method’. 3. S. pp. M. (Ed.226. S. London. pp. Porto Venere. Vol.101–117. and Holling. Agriculture. 1.R. Sage Publications Inc. Boca Raton. and Giampietro. Washington DC. 22.): Advances in Human Ecology.N. (1997) ‘Linking technology. T. Universitat Autonoma de Barcelona. 6. Ecosystems. p. European Journal of Operational Research. 3. Garcia. 109. A system approach’. Boca Raton.145–167. (Ed. Gomiero. 51. (2003) ‘Multi-objective integrated representation (putting in perspective biophysical analyses) applied to aquaculture’.315–352. Columbia University Press. Boca Raton.384. M. CRC press. pp.): Proceeding of the 3 International Workshop ‘Advances in Energy Studies: Reconsidering the Importance of Energy’. environmental loading. C. pp. M. Vol. Island Press. Ecosystems.) (1990) Emics and Etics: The Insider/outsider Debate. Goitouni.R. and Pastore.761.J. G. J.L. S. (Eds. A. pp. p. Gomiero.. M. 21–26 September 2002. Garcia. demographic pressure. M. Leclerc. (1999) ‘Multidimensional reading of the dynamics of rural intensification in China: the AMOEBA approach’... (2004) Multi-objective Integrated Representation (MOIR) as a Tool to Study and Monitoring Farming System Development and Management.org/~headlandt/ee-intro. (2001) ‘Operationalizing the concept of sustainability in agriculture: characterizing agroecosystems on a multi-criteria.. (Eds.

504. G. Hülsbergen. and Christen. Italy. J. Vol. and Munda. p. J.A. H. JRC/ISIS/MIA.. (2002) ‘Evaluating the sustainability of complex socio-environmental systems.101–109. Applied Cognitive Psychology. and Tietje. (1980) La grammatica del vedere: Saggi su percezione e gestalt. R. S. 35. pp. (Eds. pp. qualitative and fuzzy evaluation problems’. and Raiffa. (1999) Indicators for Decision-making. Kluwers Academic Publishers. Cinco experiencias de evaluación en el México rural.721–742.. R. (1999) ‘An ecosystem approach for sustainability: addressing the challenge of complexity’..360. Agriculture Ecosystems and Environment. in Italian. 31. Vol. Cuauhtémoc. S.. 4. and López-Ridaura. (1999) ‘Multiple-criteria methods for quantitative.248. S. (2003) ‘Field related organisms as possible indicators for evaluation of land use intensity’.298 T. Il Mulino. p.. original: Gestalt Psychology. Vol. 11. (1976) Decisions with Multiple Objectives: Preferences and Value Trade-Offs. St. Liveright Publishing Corporation. México. G. Regier. 10.589. (Ed. 98. H. Kauffman.709. Kanizsa. P. The MESMIS framework’.. Ecosystems. Mundi-Prensa. New York.A. T. LEEC (Laboratories of Ecotoxicology and Environmental Chemistry (I) State general laboratory of the republic of Cyprus) (2000) Integrated Monitoring & Early Warning System to Sustain Quality and Multifunctionality of Surface Waters in Cyprus for Man and the Ecosystem. (1998) ‘Ecosystems and the biosphere as complex adaptive systems’. p. Vol.S. Hofstetter. 1. Cheltenham. No. pp.it/envind/.23. .277–286.. M. H. Phoenix. Gomiero and M. (1947) La psicologia della gestalt. and Aster.185–226. (1993) Complexity: Life on the Edge of Chaos.350. in Spanish. O. available at URL: http://esl. Masera.. Hofstetter. pp.M. European Commission. Journal of Industrial Ecology. Vol. Ecosystems. p. Reading. Ecological Economics. pp. Masera.208. W. pp. Italy.184.390–405. G. Mettier. Papaja.) (2000) Sustentabilidad y sistemas campesinos. Vol. pp. pp. Martinez-Alier. (1998) Perspectives on Life Cycle Impact Assessment: A Structured Approach to Combine Models of the Technosphere. Edward Elgar. Vol. Oxford University Press. (1989) ‘Understanding charts and graphs’. London. S. P. 4. Wittmann. I-21020 Ispra (VA). J.65–99. and O’Neill J. 3. John Wiley & Sons. S.. p. Ecosphere and Valuesphere.453–461. Journal of Multi-Criteria Decision Analysis.H.97–115. O. Vol. 26. 3. Köhler. Bologna. UK. Larkin.431–436. and Francis. Müller-Wenk. Holling. New York. title translation The Grammar of Seeing: Essays on Perception and Gestalt. M. MA. Holland. Ecological Indicators. p. Levin. Vol. R. p. New York. pp. (2001) ‘Understanding the complexity of economic. and social systems’. Keeney. K. Kosslyn. in van den Bergh. O. Feltrinelli. Janssen. Special Issue: Multicriteria Analysis for Environmental Management.jrc. Vol. 2. Lewin. No.1–14. R.H. O.. (1998) ‘Weak comparability of values as foundation for ecological economics’. J. Janssen. G. Giampietro Heyer. Jesinghaus.J. López-Ridaura. J.A.): Handbook of Environmental and Resource Economics. Kay. pp. (1993) The Origins of Order: Self-organization and Selection in Evolution. pp. ecological.837–852. Munda. (1987) ‘Why a diagram is (sometimes) worth ten thousand words’. A. p. R. C. and Simon. Ch. Futures. (1999) ‘The mixing triangle: correlation and graphical decision support for LCA-based comparisons’. Braunschweig. Bologna. Addison-Wesley. (2001) ‘On the use of multi-criteria analysis in environmental impact assessment in the Netherlands’. Boyle. Cognitive Science. p. J. W. (1995) Hidden Order: How Adaptation Builds Complexity. TP 361.

p. Heidelberg. (1998) ‘Multidimensional sustainability analysis: the flag model’. J. pp. 6. (2001) The Wellbeing of Nations – A Country-by-Country Index of Quality of Life and the Environment. M. McCown. New York. (1999) ‘Conventional and land-time budget analysis of rural villages in Hubei Province. Ecological Economics.. D. (1980) From being to becoming: Time and Complexity in the Physical Science. pp. J.Graphic tools for data representation in integrated analysis 299 McConnell. Vol. .) Theory and Implementation of Economic Models for Sustainable Development. H. The World Conservation Union (IUCN). P. C. Pastore.662–677.uk/stellent/ groups/odpm_about/documents/page/odpm_about_608524. Risbey. New York. and Ravetz.378.nl/eco-indicator99/triangle. pp. (2004) ‘Social multi-criteria evaluation (SMCE): methodological foundations and operational consequences’.hcsp. in association with Rietveld. available in internet at URL: http://www. 74. Vol. p. (1990) Economics of Natural Resources and the Environment. R.. G.350. and Dillon. and Voogd.htm.net/ downloads/reports/E2001-10. and Vreeker.odpm. R.iunc.nusap. Munda.. Vol. G.319. New York. 3. NERA (National Economic Research Associates) (2002) Multi-criteria Analysis Manual. D. W.272..odpm. Critical Review in Plant Sciences. (1971) Environment.285. R. P. M.uk/stellent/groups/odpm_about/ documents/pdf/odpm_about_pdf_608524. Vol.255. Freeman & Co. and the concept of sustainable development’. New York. p.7–27. G. Munda. John Wiley & Sons. (2001) A Protocol for Assessment of Uncertainty and Strength for Emission Data. Ecological Economics.11–25. H. (1995) Multicriteria Evaluation in a Fuzzy Environment. p. pp. G. Wiley-Interscience. Vol.255–277. p.pdf and at URL: http://www. van der Sluijs. 158. R. Journal of Multi-Criteria Decision Analysis. pp. Rosen. (1993) ‘Multiple-criteria decision aid: some epistemological considerations’. Pearce.145. available in internet at URL: http://www.25. P. and Hofkes.W.J. Nijkamp. J. Prescott-Allen. Switzerland.PDF. Nijkamp. What it’s for and How to Use it.N. Columbia University Press. Prigogine. J.331. Nijkamp. Munda. p. 2. in internet at URL http://www. I.L. (2000) Essays on Life Itself. Munda.331–357.97–112. J. van Veenedaal. (1994) ‘Qualitative multicriteria evaluation for environmental management’. pp.T. ecological economics. R. Environmental Values. Munda. Ji. Power. Prescot-Allen. 2. and Li. and Society. available on internet at URL: http://www. and Turner. H.org/themes/eval/english/barom.H.htm. No. Gland. FAO. R. pp. New York. G.213–233. Rome. Columbia University Press. (1996) Indicators for Sustainable Development. and Fabrication of Life. Pretty. W. Contributions to Economics Series. Nijkamp.gov.W. European Journal of Operational Research. No. Agricultural Systems. and Rietveld.361. (2002) ‘Locating agricultural decision support systems in the troubled past and socio-technical complexity of ‘models for management’’. Origin. R.L. p. China’. G. and Ouwersloot. p. Spronk. 18. (1979) Multidimensional Spatial Data and Decision Analysis. National Academy Press. (1995) Regenerating Agriculture: Policies and Practice for Sustainability and Self-reliance. P. 10.. No..J. 3. Vol.320.M. p. pp. Physica-Verlag. P.gov. Pré Consultants (2000) Eco-indicator 99 – The Triangle Concept. Rosen. IDRC/Island Press. Vol. Giampietro. in van den Bergh.355. Odum. (Eds. Harvester Wheatsheaf.41–55. p. Kluwer. P. 33. (1997) ‘Environmental economics.K. (1991) Life itself: A Comprehensive Inquiry into the Nature. (1997) Farm Management for Asia: A System Approach. p. (2000) ‘Sustainability assessment of development scenarios: methodology and application to Thailand’. J. p.pre.

. O. Vol. Problem Solving. p. Tisdell. and Sustainable Societies. D. B. (1962) ‘The architecture of complexity’. Routledge. Sci. Nuijten. Italy.. S. O. 410 200 104. Dutch National Research Programme on Global Air Pollution and Climate Change. Portovenere. D. Vol. 698. E. pp. Petersen. IHE-Delft home page is located at URL http://www.61–76. (Eds.467–482. European Journal of Agronomy. P. B. and Clayton. in Ulgiati. J. P. S. EAAP Publication No. Funtowicz.. Vereijken. Cambridge. in Klahr. 7. Am.T. (Ed. (1981) ‘A quantitative method for description and assessment of ecosystems: the AMOEBA approach’. Segua. pp. and Ulgiati. (1993) Global Ecology: Environment Change and Social Flexibility. pp. Ifar.265–270. (1981) ‘The framing of decision and the psychology of choice’.. p. Lanhem.H. P. Wageningen University.R.225.124–129. I. Vereijken.. Udo. van Detlef.. Using the NUSAP Method. No. PhD thesis dissertation. Cheshire.. 89. p.126. Vol.91. H.. Agric. (1997) ‘Testing livestock interventions for marginal upland areas in East Java’. More than Food Production.126.L. (2001) ‘Integration of thermodynamic. in Sørensen. Hopser. p. Sucur. H. in Kuik. and Verbruggen.R. Philos. R.A. J. Lawrence Erlbaum Associated Publisher Hillsdale. (Eds. Graphics Press. S.M. Proc.. Verburg. (Ed.. SGE editoriali Padova. Marine Pollution Bulletin. H. 106. A. (1997) ‘The maintenance of natural capital: motivation and methods’. A.T. (1988) ‘The scientist as a problem solver’. P.A. Giampietro Simon. Connecticut. Tainter.): Getting down to Earth: Practical Application of Ecological Economics. and Kotovsky. Gomiero and M. and Colijn.197. Stoppato. Wageningen. Tufte. Beusen.453–458. J.A. Brown. UK. p. V. and Strategies’. Vol. in Light. 40. A. (1993) The State of the Environment of the Aegean Ionian Sea – Greece..): Philosophy and Geography I: Space.J. (1983) The Visual Display of Quantitative Information. Risbey. Washington DC. A. Corral Quintana. Tonon. F. (Eds. pp. H. The M. and Kahenman. van der Sluijs. New Jersey. H. J.375–397. and Martinez-Alier.300 T. (1993) Environmental Economics: Policies for Environmental Management and Sustainable Development. C.): Complex Information Processing: The Impact of Herbert A. 6.pdf. F. 23. p. V.. pp. B. Constraints. IHE-Delft. Place and Environmental Ethics. London. Island Press. Kluwer. Vol. Soc.H.. S. May 2000.T. de Vries.): Proceeding of the 2 International Workshop ‘Advances in Energy Studies: Exploring Supplies.. ten Brink.. (1969) The Science of Artificial. A. economic and environmental parameters for the evaluation of energy systems’.. (1990) Envisioning Information. Luchi. Graphics Press. J. A. Kloprogge. Tufte. (1996) Complexity.A. A.): Livestock Farming Systems.nusap. Connecticut. JRC Report No. S. and Smith.M..ihe. pp. The Netherlands. Simon. with Special Reference to China. Rowman & Littlefield Publisher. P. in Costanza. and Ravetz. Wageningen Pers. Science.M. F. (1992) ‘A methodic way to more sustainable farming system’. available in internet at URL: http://www. D. ten Brink. pp. and Steenstra.209–223. S. A. Simon..635–647.nl.net/workshop/report/finalrep. (2002) Uncertainty Assessment of the IMAGE/TIMER B1 CO2 Emissions Scenario. Tversky. E. M. New York. (1992) ‘The AMOEBA approach as useful tool for establishing sustainable development?’. pp. Press. London.259.235–250. p. p. K. Cheshire.143–173. 211..A. C. Simon. (Eds. J.71–87. Heuberger.240. Spash. Potting. pp. pp. J. J. Inc. (2000) Exploring the Spatial and Temporal Dynamics of Land Use. Mirandola.148.I. (1997) ‘A methodological way of prototyping integrated and ecological arable farming system (I/EAFS) in interaction with pilot farms’. S. Edward Elgar. pp. Smil. Dordrecht.): In Search of Indicators of Sustainable Development. Boulder.. Neth.

Englewood Cliffs. F.F. P. Nijkamp. Wiener.212. Zar. p. available at http://greywww.H. S. Directorate General for International Cooperation. (1948) Cybernetics. pp. New Jersey.life95cy-amoeba..pdf.uady. p. 3 Documents concerning project.mx/sitios/veterina/ofacad/ curso_protropico/material/libros/agroecologia/agroecologia. (2001) A Multicriteria Decision Support A Multicriteria Decision Support Methodology for Methodology for Evaluating Airport Valuating Airport Expansion Plans Expansion Plans. Danielson.ihe. RIVM (Research for Man and Environment). L.H. National Institute of Public Health and the Environment. 130. J.M. Wolf.kub. IHE Delf.53. New York. 12. (1999) Manual for Prototyping Integrated and Ecological Arable Farming Systems (I/EAFS) in Interaction with Pilot Farms. (1995) Recasting Alternative Agriculture as a Management Model: The Value of Adept Scaling. John Wiley.htm.nl/i_eafs/i_eafs. (1984) Biostatistical Analysis. De Zwart.nl:2080/dgr/uvatin/01005. Vol.21. p. Prentice Hall. 4 From the Proceedings of the Workshop organised by the Land and Water Development Division FAO Agriculture Department and the Research. Wefering. p.pdf. TI 2001-005/3 – Tinbergen Institute Discussion Paper.Graphic tools for data representation in integrated analysis 301 Vereijken..5–12..gov. (2000) ‘Using the AMOEBA approach to measure progress toward ecosystem sustainability within a shellfish restoration project in North Carolina’. Ecological Economics. pp. R. Hydraulic. tools and results can be found at the project website Life 95/CY/B2/CY/868 MED.H. and Ter Welle.gcw. p.718. the Dutch International Institute for Infrastructural. Copy can be requested at Drs. T. Bilthoven.cy/. and Environmental Engineering.E.157–166. Notes 1 An edition in Spanish (1999) titled Agroecologia: Bases científicas para una agricultura sustentable is freely available in internet at URL http://www.pdf. Extension and Training Division FAO Sustainable Development Department 25–26 January 1996. available in internet at URL: http://www. P.M. 2nd ed. Verhagen.11. Ecological Modelling. The Netherlands. and Allen. N. a summary is available in internet at URL: http://www. Wageningen. C.A.131–155. N. Stichting Cereales. Vol. The Netherlands.318. J. p. The document is available in internet at URL http://www. and White.nl/he/dicea/monitor/ aegean. . 2 The manual has been prepared by order of the Netherlands Ministry of Foreign Affairs. (1999) Netcoast Open Learner – Coastal Monitoring: Amoeba of the Aegean Sea. pp. Vreeker.

Sign up to vote on this title
UsefulNot useful