Modelling qualitative knowledge for strategic river management: necessity, feasibility and utility

In decision making processes on strategic river management, use of models is not as great as the research efforts in the field of model application might suggest they could be. Both the fact that the development of many models remains restricted to readily available data and pre-existing models, alongside a failure to address uncertainties, are regarded as symptomatic of this. Both are addressed. A study of the Integrated Explorative Study of the Dutch river Meuse (in Dutch: Integrale Verkenningen Maas, IVM) shows how stakeholders in group decision making processes use criteria which structurally differ from those addressed by a model utilized in the same process; the first using more abstract, the second more concrete criteria to address similar problems. Construal level theory, originating from consumer psychology, explains how the difference between these two can influence the decision making process. Addressing more abstract criteria in models could help bridging the gap between models and their users. These abstract criteria are often based on qualitative knowledge. Fuzzy logic can be used to quantify and incorporate qualitative knowledge in models. The uncertainty in the outputs of such models, due to uncertainty in inputs, in parameterization of the sets and due to overlap between sets and the shape of sets, can be quantified. Restrictions to the application of fuzzy logic in this context are the requirements of unambiguously defined relations and corresponding qualitative classifications. Fuzzy logic is relatively easy to apply to criteria which are already described in classes between which gradual transitions occur, such as in our case agriculture suitability. The qualitative, quantitative and uncertainty information were used to explore how the different types of information representation affect decision making. A query was offered to respondents from river engineering and river management backgrounds. Response shows that particularly the uncertainty information has a large influence on the decisions made; the effect of quantifying qualitative knowledge turned out not statistically significant. Concerning the uncertainty outcomes, respondents independently adopted an uncertainty strategy in which the ‘risk of obtaining a negative outcome’ across all decision criteria is the smallest.