Uncertainty in integrated assessment modelling: A labyrinthic path

Abstract

Integrated Assessment is the practice of combining different strands of knowledge to accurately represent and analyse real word problems of interest to decision-makers. Since these problems rarely observe disciplinary boundaries, Integrated Assessment usually involves interdisciplinary research. However, what distinguishes Integrated Assessment from interdisciplinary research is its policy dimension, aiming to inform decision-makers on the complexity of real world problems. Unfortunately, the body of existing disciplinary knowledge is often insufficient for the construction of an accurate representation of real world problems. Integrated Assessment offers a systematic approach to identification of the gaps in disciplinary knowledge that have often frustrated policy analysis in the past. Thus, Integrated Assessment has increasingly been the source of critical questions and new directions of research in the disciplinary sciences. Integrated Assessment is particularly useful for analysis of real world problems that are complex, operate at different levels in time and space, are immersed in uncertainty and for which the stakes are high. Because there are no simple solutions to these complex problems facing humankind, Integrated Assessment aims at conveying innovative and sometimes counterintuitive insights into the issues at hand rather than ready-made solutions. Portraying and translating real world problems can be done from a plurality of perspectives. There is no one “right” way to represent and analyse the world, therefore a diversity of methods and approaches to Integrated Assessment are needed, ranging from model-based methods to participatory methods [22,29]. Generally, these methods are, in varying degrees, in their relative infancy. The currently most widely used method of performing Integrated Assessment is modelling. Integrated Assessment models are frameworks to organize and structure various pieces of recent scientific disciplinary knowledge. A key issue in Integrated Assessment (IA) modelling is uncertainty due to various reasons. First of all IA modelling is confronted with the inherent uncertainty and lack of knowledge that the disciplinary sciences face. Secondly, IA models have to deal with a variety of types and sources of uncertainty that have to be structured and combined in one way or another. And finally, IA models are prone to a cumulation of uncertainties, because of their ambition to cover the whole cause–effect chain of a particular real world problem. This all makes uncertainty one of the most problematic but also one of the most challenging issues in the field of IA modelling. This paper therefore focuses on the laborious relation between uncertainty and IA modelling. After a description of what IA models are and where they can be used for, the issue of uncertainty is raised and how IA models struggle with it. One possible way out is presented in terms of a pluralistic approach towards the management of uncertainties in IA modelling.