Defining assessment projects and scenarios for policy support: Use of ontology in Integrated Assessment and Modelling

Abstract

Integrated Assessment and Modelling (IAM) provides an interdisciplinary approach to support ex-ante decision-making by combining quantitative models representing different systems and scales into a framework for integrated assessment. Scenarios in IAM are developed in the interaction between scientists and stakeholders to explore possible pathways of future development. As IAM typically combines models from different disciplines, there is a clear need for a consistent definition and implementation of scenarios across models, policy problems and scales. This paper presents such a unified conceptualization for scenario and assessment projects. We demonstrate the use of common ontologies in building this unified conceptualization, e.g. a common ontology on assessment projects and scenarios. The common ontology and the process of ontology engineering are used in a case study, which refers to the development of SEAMLESS-IF, an integrated modelling framework to assess agricultural and environmental policy options as to their contribution to sustainable development. The presented common ontology on assessment projects and scenarios can be reused by IAM consortia and if required, adapted by using the process of ontology engineering as proposed in this paper.

Introduction

Integrated Assessment and Modelling (IAM) is increasingly used to assess the impacts of policies, technologies or societal trends on the environmental, economic and social sustainability of systems (Harris, 2002, Parker et al., 2002, Oxley and ApSimon, 2007, Hinkel, 2009). Prominent examples are the assessment of climate change impacts (Weyant et al., 1996, Cohen, 1997, Warren et al., 2008) and the assessment of quality and allocation effects in water resource management (Turner et al., 2001, Letcher et al., 2007, Ticehurst et al., 2007). Integrated assessment is defined by Rotmans and Asselt (1996) as an interdisciplinary and participatory process of combining, interpreting and communicating knowledge from diverse scientific disciplines to allow a better understanding of complex phenomena. IAM is then a methodology to combine several quantitative models representing different systems and scales into a framework for integrated assessment (Parker et al., 2002). Consequently IAM can cover several organisational and spatio-temporal scales to provide quantitative assessment of impacts on sustainable development.

Core features of any IA are the integration among disciplines and between scientists and stakeholders (Rotmans, 1998, Parker et al., 2002). Scenario analysis is an important technique in integrated assessment (Rotmans, 1998), where scenarios are developed and used in the interaction between scientists and stakeholders to anticipate and to explore possible futures and to assess potential consequences of different strategies into the future. The literature provides many different definitions of the concept scenario. For example, Rotmans (1998) defines scenarios as ‘archetypal descriptions of alternative images of the future, created from mental maps or models that reflect different perspectives on past, present and future developments,’ while Parry and Carter (1998) define a scenario as ‘a coherent, internally consistent and plausible description of a possible future state of the world.’ In strategic business planning, where scenarios are often used as planning tool, scenarios are defined (according to Schoemaker, 1993) as ‘focused descriptions of fundamentally different futures presented in a coherent script or narrative.’ Peterson et al. (2003) provide a definition of scenario which is closer to modelling, i.e. ‘as variation in the assumptions used to create models.’

Next to a wide range of definitions for scenarios, also different classifications and typologies of scenarios exist (Rotmans, 1998, Greeuw et al., 2000, Alcamo, 2001, Van Notten et al., 2003, Borjeson et al., 2006): forecasting vs. backcasting scenarios, descriptive vs. normative scenarios, quantitative vs. qualitative scenarios, trend vs. peripheral scenarios, baseline vs. policy vs. business-as-usual scenarios and exploratory vs. anticipatory scenarios. A wide diversity of terms are associated with scenarios, such as indicators, driving forces, time horizon, time steps, storyline or narrative, processes, states, events, consequences and actions. It is not clear how these classifications and terms relate to each other and how they are used in constructing scenarios for IA.

Confusion and misunderstanding are particularly high when it comes to the implementation of scenarios. A researcher who is working in an IAM team, will be confronted with different types of stakeholders and scientists, with the latter covering a wide variety of disciplines and experiences. Each scientist will have a specific understanding of the concept scenario which is not consistent across disciplines and models. Discussions among scientists from different disciplinary domains and stakeholders are likely to result either i) in developing a ‘container’ term for scenario which serves as the magical solution whenever researchers are unclear about the way forward, or ii) in lengthy discussions on the meaning of scenario without arriving at any conclusion acceptable to the whole group. Again, the critical issue is that different models and policy problems have a specific implementation of scenarios targeted at that specific model or policy problem. There is a need for a clear set of rules and protocols with respect to scenarios in IA, as concluded by Rotmans and Asselt (1996), to avoid the danger of in-transparent, inconsistent, narrowly defined and ad-hoc setting of parameters (Rotmans, 1998, Van Asselt and Rotmans, 2002).

This paper considers a case study of achieving consensus on scenario definition in an IAM consortium, System for Environmental and Agricultural Modelling; Linking European Science and Society (SEAMLESS) (Van Ittersum et al., 2008). It provides a computerized framework (SEAMLESS-IF) to assess the sustainability of agricultural systems in the European Union at multiple scales. The SEAMLESS consortium includes 30 institutions and more than 100 researchers from agronomy, economics, landscape ecology, social science, environmental science and computer science with dissimilar research background, leading to many different views on the meaning of scenario and its implications for the computerized integrated framework (SEAMLESS-IF). For example, biophysical simulation models (Van Ittersum and Donatelli, 2003) used for climate change impact assessment often apply the SRES scenarios framework (IPCC, 2000). In contrast, in a market model (Britz et al., 2007) a scenario typically refers to a policy that might be implemented in the future and that affects the market.

This paper proposes a unified structured view for model-based scenario and assessment projects and a process of arriving at this result within a large community of researchers in a consortium. We demonstrate the use of common ontologies (see Section 2 for explanation) in building this shared conceptualization through a case study. This paper describes our experiences in the challenging task of arriving at a shared conceptualization among researchers from different disciplines with dissimilar education and research experiences. We suggest that the process and the methods used are reusable for different integrated assessment tools or consortia developing such tools.

In Section 2, the theory behind common ontologies and the process of ontology engineering will be explained. Also, our case study based on the SEAMLESS consortium is introduced. In Section 3, the developed common concept on scenario and assessment projects is presented, including one fictitious example of the use of the common concept in an integrated assessment project at the regional scale. The common concept is discussed in Section 4. In Section 5 we address our main findings as to the unified structured view on scenarios and assessment projects that we propose in this paper. Throughout the paper, we list some of the lessons we learned in our exercise to achieve this common understanding.