**Using Fuzzy Cognitive Mapping in Environmental Decision Making and Management: A Methodological Primer and an Application**

Elpiniki Papageorgiou1 and Areti Kontogianni2

*1Technological Educational Institution of Lamia, Department of Informatics and Computer Technology, Lamia, 2University of Aegean, Department of Marine Sciences, Mytilini, Lesvos, Greece* 

## **1. Introduction**

426 International Perspectives on Global Environmental Change

Lakshmi Kumar, T.V., Humberto Barbosa, K. Koteswara Rao & Emily Prabha Jothi, 2011,

MEA (Millennium Ecosystem Assessment)**.** Ecosystem and Human Well-Being: Desertification Synthesis: Island. Press World Resources Institute, Washington, DC, 2005b. Namias, J (1972) Influence of northern hemisphere general circulation on drought in

Nobre, P., Shukla, J (1996) Variations of sea surface temperature, wind stress and rainfall over the tropical Atlantic and South America, *Journal of Climate*, 9, 2464-2479. Nicholson S (2005) On the question of the "recovery" of the rains in the West African Sahel,

Nicholson, S. E & Farrar, T. J (1994) The influence of soil type and the relationships between

Pathak P.N., Rao B.M and Potdar M.B (1993) Large area soil moisture estimation using INSAT – VHRR infrared data, *Indian Journal of Radio & Space Physics*, 22, 156-164. Rao B M., Thapliyal PK., Manikiam B., Avinash Dwivedi (2001) Large scale soil moisture

Reynolds, J. F. et al. Natural and human dimensions of land degradation in drylands: causes

Rosenzweig, C., Parry, M.L (1994) Potential impacts of climate change on world food

S. Nagai, K.Ichii., H.Morimoto (2007) Interannual variations in vegetation activities and climate variability caused by ENSO in tropical rainforests, 28(6)1285-1297. Sarma, A.A.L.N & Lakshmi Kumar, T.V (2006) Studies on crop growing period and DVI in

Sarma A.A.L.N & T.V. Lakshmi Kumar (2007), Rainfall and soil wetness response to ENSO

Sarma A.A.L.N & T.V. Lakshmi Kumar (2006) Studies on agroclimatic elements and soil wetness estimation from MSMR data, *Journal of Agrometeorology*, 8 (1), 19-27. Sarma, A.A.L.N., Padma Kumari, B., Srinivas S (1999) Studies on hydrological extremes –

Sarma A.A.L.N., Lakshmi Kumar, T.V., Koteswara Rao, K (2010). Extreme value analysis of

Thapliyal P.K., Rao B.M., Pal P.K and Das H P, 2003 Potential of IRS-P4 microwave radiometer data for soil moisture estimation over India, *Mausam*, 54, 277 – 286.

Verdin, J. P., C. Funk, G. Senay, and R. Choularton (2005) Climate science and famine early

White, M. A., Thornton, O. E., and Running, S. W (1983) A continental phenology model for

Thoronthwaite C W & Mather J, The water balance , Publ Clim USA, 8, 1955, 1.

warning, *Philos. Trans. R. Soc. London*, 360(1463), 2155–2168.

Sarma A.A.L.N & T.V. Lakshmi Kumar, 10/4/375, ISRO RESPOND REPORT, 2005.

NDVI, rainfall and soil moisture in semiarid Botswana, *Remote Sensing. Environment.*

estimation using microwave radiometer data, *Journal of Agrometeorology*, 3(1&2), 179

and consequences. In: Canadell, D. P. L. F. P. (ed.), Terrestrial ecosystems in a

relation to water balance components, *Indian Journal of Radio & Space Physics*, 35,

and soil wetness estimation from remote sensing data, *Journal of Indian Geophysical* 

summer monsoon rainfall over Andhra Pradesh, *Indian Journal of Radio & Space* 

monitoring vegetation responses to interannual climate variability. *Global* 

Science & Technology, Accepted for publication.

northeast Brazil, *Tellus*, 4, 336-342.

50, 107-120.

– 187.

424-434.

*Union*, 11(2), 91-100.

*Physics*, 39, 32-38.

*Biogeochem. Cycles*, 11, 217-234.

*Journal of Arid Environments*,63, 615–641.

changing world. *Springer*, 247-258, 2007.

ENSO signal, IAHS publication, No.255, 73-80.

supply. *Nature*, 367,133-138.

*Studies on the freuency of extreme weather events over India*, Journal of Agriculture

Widespread concerns over the integrity of natural ecosystems worldwide have initiated numerous attempts at developing new tools of monitoring present conditions, assessing future risks and visualizing alternative futures. Reports on the 'state of the world' abound and so do policy proposals and sustainability strategies. Amidst this plenty of ideas, our ability to reverse the trend and secure a safe, minimum stock of valuable natural capital seems counterproductive. A better understanding of ecosystem dynamics at both the quantitative (biochemical cycling) as well as the qualitative (ecological structure of food webs) levels, without artificial divisions between them, is needed. We also need to understand better the institutional failures leading to a growing number of 'tragedies of the commons'.

To tackle these challenges appropriately, current environmental management strategies need to 'navigate' through an apparent tension: On the one hand they must meet the demand for scientific knowledge-based policy, expressed under the motto 'science speaks to policy'. On the other hand, the very same strategies urge for stakeholder involvement and sponsor initiatives to elicit lay-people attitudes, beliefs and visions for the future. This tension seems to reflect the ever lasting stand-off of bottom up and top down approaches.

The motivation for this chapter comes from the authors' uneasiness with the present methodological arsenal in the domain of environmental stakeholder analysis. Previous research on non-market valuation of environmental assets has shown the importance of complementing the neoclassical microeconomic framework of choice in stated preference surveys with qualitative - both ex ante and ex post – analysis of individual mental processes, perceptions and beliefs (Kontogianni et al, 2001, 2005, 2008). Especially applications of contingent valuation have benefited from in-depth interviews and focus groups conducted ex ante in order for the researcher to understand the cultural, social and psychological background of choices elicited through structured interviews. (Desvousges & Smith 1988, Brouwer, 1999). In spite though of the importance of stakeholder qualitative analysis in

Using Fuzzy Cognitive Mapping in Environmental

specific disadvantages as Kok (2009) overview them.

groups of stakeholders (Ozesmi & Ozesmi, 2004).

**3. The structure of Fuzzy Cognitive Maps** 

management in precision agriculture (Papageorgiou et al., 2009, 2010).

symbolic representation for the description and modeling of a system.

Decision Making and Management: A Methodological Primer and an Application 429

tools by applying fuzzy cognitive mapping (FCM) to the exploitation of local knowledge. FCM fits the requirements stated above better than any of the other conceptual modelling techniques analyzed here. Most other methods are either too difficult for the type of stakeholders we are aiming for, or take too much time. Yet, FCMs have their own set of

In the case of stakeholders' analysis for ecological modeling and environmental management, the FCMs have found a good number of applications. At first, Hobbs et al., 2002 applied FCM as a tool to define management objectives for complex ecosystems (Hobbs et al., 2002). Next, Ozesmi and Ozesmi (2003, 2004) proposed a multi-step FCM and participatory approach of Stakeholder Group Analysis in Uluabat Lake, Turkey, for ecosystem observation. The multi-step fuzzy cognitive mapping approach analyzes how people perceive a system, and compare and contrasts the perceptions of different people or

After the pioneering work of Ozesmi & Ozesmi (2003, 2004), in environmental and ecological management topics, other researchers followed with more implementations of FCMs in this area. FCMs have been employed in a number of studies including a FCM for rapid stakeholder and conflict assessment for natural resource management (Hjortsø et al. 2005; Robson & Kant, 2007), a FCM for modelling a generic shallow lake ecosystem by augmenting the individual cognitive maps (Tan & Ozesmi, 2006), FCM for predicting the effects of perturbations on ecological communities, thus to control on the fledging rate of an endangered New Zealand bird (Ramsey & Vetman, 2005), FCM for assessing local knowledge use in agroforestry management (Isaac et al., 2009), FCM for modelling of interactions among sustainability components of an agro-ecosystem using local knowledge (Rajaran & Das, 2009), FCM for predicting modelling a New Zealand dryland ecosystem to anticipate pest management outcomes (Ramsey & Norbury, 2009), FCM for cotton yield

Fuzzy Cognitive Mapping methodology is a symbolic representation for the description and modeling of complex systems. Fuzzy Cognitive Maps (FCMs) describe different aspects of the behavior of a complex system in terms of concepts. Each concept represents a state or a characteristic of the system and interacts with each other showing the dynamics of the system. FCMs have been introduced by Kosko, (1986) as signed directed graphs for representing causal reasoning and computational inference processing, exploiting a

In fact, FCM could be regarded as a combination of Fuzzy Logic and Neural Networks (Kosko, 1992). Graphically, FCM seems to be an oriented graph with feedback, consisting of nodes and weighted arcs. Nodes of the graph stand for the concepts that are used to describe the behavior of the system, connected by signed and weighted arcs representing the causal relationships that exist between the concepts (see Figure 1). It must be mentioned that all the values in the graph are fuzzy, so concepts take values in the range between [0,1] and the weights of the arcs are in the interval [-1,1]. Observing this graphical representation it becomes clear which concept influences other concepts by showing the interconnections between them. Moreover, FCM allows updating the construction of the graph, such as the adding or deleting of an interconnection or a concept. FCMs are used to represent both

stated preferences valuation techniques, we still lack a coherent, standardized approach to analyze environmental perceptions and beliefs. The need to fill the gap becomes apparent when we recognize the fact that the way non-experts articulate complex relationships, such as those governing marine ecosystem functions, have their own special weight in influencing policy design and implementation: they transcend the fact/value divide (senso Putnam 1985) and offer valuable insights on the ways cause and effect relationships in nature are perceived (Karageorgis et al., 2006).

Fuzzy Cognitive Mapping (FCM) was thus selected as a suitable method for semiqualitative analysis to achieve our research goal. In this chapter we introduce the reader to the concept of Fuzzy Cognitive Maps (FCMs) and their theoretical background. In section 2 we summarize the state-of-the-art in qualitative, stakeholder analysis for environmental management. We then present the structure of FCMs (section 3) and the analytical use of graph theory in defining relevant indices (section 4). We proceed with the development of FCMs (section 5) and the FMC inference and simulation processes (section 6). After presenting the theoretical structure, the practical steps involved in the design and implementation of a FCM exercise are codified (section 7). We then illustrate the concepts discussed so far with a practical application implemented by the authors in The Black Sea (section 8) before we summarize and conclude in section 9.

### **2. The many facets of stakeholder analysis in environmental management**

Integrated approaches to environmental planning with proper stakeholder involvement offer a possible way forward. Such an approach needs to facilitate communication within multidisciplinary research teams; it needs to recognize the functional continuity from watersheds to the coasts to the open sea, thereby helping to locate the scale of intervention less on the base of traditional jurisdictions and more towards appropriate ecosystem scales. Last but not least, it must encompass participatory management schemes which promise a substantive change in the exploitation of local knowledge. By enhancing stakeholder involvement, participatory management strengthens policy relevance, diminishes uncertainties, improves monitoring and raises enforcement rates (NRC 1996, OECD 2005). Participatory (or deliberative) approaches to environmental management are usually grouped under the general term of stakeholder analysis (Grimble and Wellard 1997, Bryson 2004, Reed et al., 2009). Stakeholder analysis in turn can be divided into what we opt to call macro-stakeholder and micro-stakeholder analysis. The former category includes all those qualitative approaches that refer to the interaction of social groups and their dynamics: social networks analysis (Scott, 2000, Carrington et al 2005, Turnpenny et al., 2005), analysis of conflicts (Howard, 1989, Hjortso et al. 2005, 2010; Stoney & Winstanley, 2001), and actor analysis (Hermans, 2008). The latter category refers to qualitative or semi-quantitative approaches, which explore individual perceptions, values and attitudes. These include: fuzzy cognitive mapping of social perceptions and values (Bots et al., 2000, Stone 2002), perceptions mapping (Bots, 2007), mind mapping (Buzan, 1993), concept mapping (Novak, 1993), focus groups and in-depth interviews.

Approaches in stakeholder analysis as described above share some common characteristics: they are 'eclectic but pragmatic' approaches with varying degree of sophistication, requiring in average a low in-depth academic investigation, but able to manipulate a vast quantity of soft information. Their strength lies primarily with thinking about problems than solving them. The present paper aims at contributing to a refinement of participatory management

stated preferences valuation techniques, we still lack a coherent, standardized approach to analyze environmental perceptions and beliefs. The need to fill the gap becomes apparent when we recognize the fact that the way non-experts articulate complex relationships, such as those governing marine ecosystem functions, have their own special weight in influencing policy design and implementation: they transcend the fact/value divide (senso Putnam 1985) and offer valuable insights on the ways cause and effect relationships in

Fuzzy Cognitive Mapping (FCM) was thus selected as a suitable method for semiqualitative analysis to achieve our research goal. In this chapter we introduce the reader to the concept of Fuzzy Cognitive Maps (FCMs) and their theoretical background. In section 2 we summarize the state-of-the-art in qualitative, stakeholder analysis for environmental management. We then present the structure of FCMs (section 3) and the analytical use of graph theory in defining relevant indices (section 4). We proceed with the development of FCMs (section 5) and the FMC inference and simulation processes (section 6). After presenting the theoretical structure, the practical steps involved in the design and implementation of a FCM exercise are codified (section 7). We then illustrate the concepts discussed so far with a practical application implemented by the authors in The Black Sea

**2. The many facets of stakeholder analysis in environmental management** 

Integrated approaches to environmental planning with proper stakeholder involvement offer a possible way forward. Such an approach needs to facilitate communication within multidisciplinary research teams; it needs to recognize the functional continuity from watersheds to the coasts to the open sea, thereby helping to locate the scale of intervention less on the base of traditional jurisdictions and more towards appropriate ecosystem scales. Last but not least, it must encompass participatory management schemes which promise a substantive change in the exploitation of local knowledge. By enhancing stakeholder involvement, participatory management strengthens policy relevance, diminishes uncertainties, improves monitoring and raises enforcement rates (NRC 1996, OECD 2005). Participatory (or deliberative) approaches to environmental management are usually grouped under the general term of stakeholder analysis (Grimble and Wellard 1997, Bryson 2004, Reed et al., 2009). Stakeholder analysis in turn can be divided into what we opt to call macro-stakeholder and micro-stakeholder analysis. The former category includes all those qualitative approaches that refer to the interaction of social groups and their dynamics: social networks analysis (Scott, 2000, Carrington et al 2005, Turnpenny et al., 2005), analysis of conflicts (Howard, 1989, Hjortso et al. 2005, 2010; Stoney & Winstanley, 2001), and actor analysis (Hermans, 2008). The latter category refers to qualitative or semi-quantitative approaches, which explore individual perceptions, values and attitudes. These include: fuzzy cognitive mapping of social perceptions and values (Bots et al., 2000, Stone 2002), perceptions mapping (Bots, 2007), mind mapping (Buzan, 1993), concept mapping (Novak,

Approaches in stakeholder analysis as described above share some common characteristics: they are 'eclectic but pragmatic' approaches with varying degree of sophistication, requiring in average a low in-depth academic investigation, but able to manipulate a vast quantity of soft information. Their strength lies primarily with thinking about problems than solving them. The present paper aims at contributing to a refinement of participatory management

nature are perceived (Karageorgis et al., 2006).

(section 8) before we summarize and conclude in section 9.

1993), focus groups and in-depth interviews.

tools by applying fuzzy cognitive mapping (FCM) to the exploitation of local knowledge. FCM fits the requirements stated above better than any of the other conceptual modelling techniques analyzed here. Most other methods are either too difficult for the type of stakeholders we are aiming for, or take too much time. Yet, FCMs have their own set of specific disadvantages as Kok (2009) overview them.

In the case of stakeholders' analysis for ecological modeling and environmental management, the FCMs have found a good number of applications. At first, Hobbs et al., 2002 applied FCM as a tool to define management objectives for complex ecosystems (Hobbs et al., 2002). Next, Ozesmi and Ozesmi (2003, 2004) proposed a multi-step FCM and participatory approach of Stakeholder Group Analysis in Uluabat Lake, Turkey, for ecosystem observation. The multi-step fuzzy cognitive mapping approach analyzes how people perceive a system, and compare and contrasts the perceptions of different people or groups of stakeholders (Ozesmi & Ozesmi, 2004).

After the pioneering work of Ozesmi & Ozesmi (2003, 2004), in environmental and ecological management topics, other researchers followed with more implementations of FCMs in this area. FCMs have been employed in a number of studies including a FCM for rapid stakeholder and conflict assessment for natural resource management (Hjortsø et al. 2005; Robson & Kant, 2007), a FCM for modelling a generic shallow lake ecosystem by augmenting the individual cognitive maps (Tan & Ozesmi, 2006), FCM for predicting the effects of perturbations on ecological communities, thus to control on the fledging rate of an endangered New Zealand bird (Ramsey & Vetman, 2005), FCM for assessing local knowledge use in agroforestry management (Isaac et al., 2009), FCM for modelling of interactions among sustainability components of an agro-ecosystem using local knowledge (Rajaran & Das, 2009), FCM for predicting modelling a New Zealand dryland ecosystem to anticipate pest management outcomes (Ramsey & Norbury, 2009), FCM for cotton yield management in precision agriculture (Papageorgiou et al., 2009, 2010).
