Introduction

This document explains the multiple criteria decision methods implemented in the mDSS5, a generic DSS developed to assist water authorities in the management of water resources. The software was originally developed in the context of the project MULINO (MULti-sectoral, INtegrated and Operational Decision Support System for Sustainable Use of Water Resources at the Catchment Scale) and further developed and applied with a contribution of several other projects, including DSS-GUIDE, TRANSCAT, NOSTRUM-DSS, NEWATER and BRAHMATWIN.

The mDSS5 is particularly useful when applied together with other tools such as stakeholders’ analysis or problem structuring methods. We apply the software to a broader analytical framework which we refer to as NetSyMod. This framework is the result of several years of research in the field of environmental evaluations and decision-making carried out at FEEM within the Natural Resources Management Research Programme. It consists of a suite of tools aimed at facilitating the involvement of stakeholders and experts in environmental decision-making processes.

The main components of the framework are:

• Identification of all potential stakeholders/experts affected by the policy decision under examination. The process proposed to fulfil this objective is a simple approach based on the organization of brainstorming meetings and the use of snowball techniques.
• Social Network Analysis (SNA) which aims to assess the reciprocal relationships among actors within their local social networks. Characterizing the power structure prevalent in the selected group of actors, SNA should ensure that the participatory modeling and/or planning process is not hijacked by powerful groups, but rather, it is truly representative of the whole sample – and population – of interested parties.
• Creative System Modelling (CSM) which provides means for facilitating the process of participatory modeling and, more specifically, for eliciting knowledge and preferences from actors. The key actors chosen in the previous steps of the NetSyMod approach will take part in a participatory workshop during which Cognitive Mapping techniques most suitable for the specific case will be applied.
• Analysis of Options, in which the participatory approach is brought in the field of decision support envisaging the use of specific computer support tools.

In the subsequent section of this chapter, these components are briefly described. For an in-depth description, see Giupponi et al. (2007). To continue with the mDSS5 decision methods, go to Chapter 2.

1.1     ACTORS’ ANALYSIS

This initial phase identifies all potential stakeholders/experts involved or affected by the decision under investigation and singles out those who should take an active part in the decision-making process.

First of all, it is necessary to identify all potential stakeholders/experts involved in, or affected by, the decision to be undertaken. Within the NetSyMoD framework, a task force group is set up for this purpose, which, through a combination of brainstorming meetings and a modified snowball sampling technique, carries out this task.

When all the relevant actors have been identified, a Social Network Analysis is undertaken, with the aim of assessing the reciprocal relationship among actors. Through the use of questionnaires and interviews, the SNA will allow the identification of key actors, the assessment of the power structure among the actors, and the characterization of their roles and position with respect to the decision to be taken.

SNA ensures that the participatory modeling and/or planning process is not hijacked by powerful groups, but rather it is truly representative of the whole spectrum of interests and positions. There are thus three main outputs from the SNA phase, which will be an input into the preparatory phase for the CSM workshop.

1. A list of key stakeholders/experts to be involved in the next phases of NetSyMoD. This will limit the number of participants to a manageable size, and ensure that no important actors are left out of the exercise.
2. The analysis of power will highlight potentially problematic actors and relations, whom the facilitator will need to actively manage during the creative system modeling workshop.
3. A conflict analysis on the basis of the position and roles of actors within the network, with the purpose of identifying key alleys and/or opponents, and actors who are opinion setters.

1.2     PROBLEM ANALYSIS

In this phase, the problem (or conflict) at hand is scrutinized from various perspectives and viewpoints. The environment in which the problem is embedded is explored and the relevant factors identified.

The problems faced by natural resource managers are complex and their drivers interwoven. It is necessary to identify the most relevant aspects, by focusing on which the major changes can be attained. The exploration of the problem includes analyses of legal and institutional frameworks, as well as the economy on various spatial levels and the state of the environment. Future development of main drivers and pressures are simulated using models under alternative scenarios. 

Different stakeholders (identified in the previous step – Actor analysis) hold different perceptions and beliefs about what are the causes of the problem or how it should be tackled. Different techniques have been developed to surface tacit knowledge and deeply held beliefs, including conflict assessment, problem structuring methods, and discourse analysis. The individual perspectives are further elaborated in the next step (Creative system modeling) to facilitate collective learning and shared (agreed) boundaries of the problem.

The problem analysis phase typically ends with

1. A list of the most relevant drivers governing the perception of the problem at hand
2. Sketch of cause-effect relations between various drivers, identified and explored using multiple methods and models
3. A set of scenarios regarding the future development of the main drivers and cause-effect relations
4. An extensive list of indicators against which the performance of policy measures should be measured

1.3     CREATIVE SYSTEM MODELLING

A shared model of reality is needed for the correct evaluation of policy options. Creative System Modelling (CSM) techniques facilitate the process of participatory modeling and elicitation of knowledge and preferences from actors, thus building a common understanding of the problem.

The key actors identified in the previous step will take part in a participatory workshop, during which cognitive mapping techniques (such as Hodgson’s hexagon method or a revised Delphi technique) will be used to develop a shared model of the decision problem.

The CSM workshop can have two main aims, depending on the case at hand:

1. building a shared model of the problem, based on cause-effects chains and using the DPSIR conceptual model (Driving Force, Pressure, State, Impact, Response); or
2. developing shared scenarios, depicting the potential evolutions of the system over time, or under different policies.

The CSM will also serve the purpose of identifying shared evaluation criteria and eliciting individual and group weights, necessary for the evaluation of policy options through multicriteria analysis.

Creative system modeling provides not only a common ground for mutual understanding among the parties involved but also a scientifically sound basis for the development of effective decision support systems (DSSs).

The cognitive map of the decision problem, or the related scenarios, will be the basis for the analysis of options:

• the shared mental maps elicited at the CSM workshop will be the underlying modeling framework for tailoring mDSS to the specific needs;
• the workshop will provide qualitative and/or quantitative indicators to be used in the choice phase in mDSS;
• the workshop may also lead to a quantitative assessment of these indicators, in addition to their identification.

1.4     DSS DESIGN

In this phase, numerous tools and information (knowledge) produced in previous steps are assembled into a toolbox or framework. This is necessary to manage the information flow between various process phases, including exchange, transformation, integration, validation and documentation of gathered knowledge.

Many of the previous analyses employ computer-based tools such as databases (and data management systems), visualization components, and simulation models. Different tools are frequently assembled into comprehensive Decision Support Systems, normally employing various interconnected and adapted components, controlled by a user interface. This phase addresses all activities related to the development of interoperable and useable software components; and the collection of well-documented and easily exchangeable data sets (including spatial data and time series).

• Seamless data flow between various tools and software component
• User interface which guides the user through various stages of the NetSyMod process
• Quality assurance regarding the integration of different components
• Documentation and report facilities which explain the process and facilitate the interpretation of results

1.5     POLICY EVALUATION

Policy evaluation consists of choosing one (or more) policy measure from a set of mutually exclusive alternatives, or producing their complete ranking. Numerous methods and techniques have been developed in decision theory to make explicit (transparent) value judgments and assess the extent to which different policies contributed to achieving the pursued goals and objectives. 

Decision models (DM) result from the systematic exploration and negotiation of a ‘problem’, including its existence, boundaries and structure. DMs comprise alternative courses of action (policies or policy measures); decision goals – translated into more tangible evaluation criteria – against which the policies are weighed; and preferences, which describe how well the policies satisfy the objectives.

There are normally several candidate policies; for example, high nitrate pollution can be tackled by introducing financial incentives, changing nutrient management in farms, protecting littoral vegetation and favoring phytodepuration, or improving the effectiveness of wastewater treatment plants, WWTP). Binary (yes/no) choices, such as whether to adhere to the Kyoto Protocol for reducing greenhouse gas emissions are frequently indicative of escalating conflicts due to incommensurable ethical principles, values and interests. Goals may refer to competing targets, e.g. macro-economic developments vs. social impact; favoring different policies so that no single option outperforms all others. In these situations, decision makers may be a priory uncertain (undecided) about what policy action is most appropriate. This indecisiveness is a result of the diversity of decision outcomes, which are not uniformly distributed in space and time (e.g. different policy impacts on upstream vs. downstream water users; WWTP extensions may have an earlier impact on nitrate concentration than land use changes) or the values attached to them. Uncertainty in the outcomes of a choice poses yet another challenge for decision making.

The trade-offs or preferences are value judgments, which are frequently not observable and must be revealed or approximated. Such uncovered preferences are context-specific and depend on the description and framing of a problem, and how the questions are formulated. For example, to assess the environmental costs of irrigation, one must consider the value of wetlands and riverine ecosystems deprived of water abstraction. These values, regardless of whether they are in monetary terms or relative utility, may be difficult to approximate as the results depend on the respondents’ prior knowledge or on what they think others would approve.  In situations involving uncertainty, preferences are formed over probabilities of possible outcomes of the policies and integrated into the decision model. These preferences embody attitudes towards risk (risk aversion vs. risk seeking vs. risk neutrality), defined according to the value individuals attach to the uncertain outcomes of a decision. 

Decision methods help to avoid inconsistencies underlying judgment and choice, and make decisions more compatible with normative axioms of rationality. Furthermore, if combined with deliberative techniques, decision methods render policy processes transparent and inform the perspectives or viewpoints of all actors. This is translated into a higher acceptance of the policies.