Uncertainty characterizes a flood risk system by definition: from forecasting meteorological and hydrological conditions to predicting socio-economic development. Uncertain is also the way river-dike interactions take place (e.g. upstream-downstream hydrodynamic interactions), whose understanding and evaluation is of paramount importance in large-scale flood risk management. This latter, system-related uncertainties manifest in many ways: What will be the hydraulic load at one location, given a breach at a location upstream? Which dikes will fail? And when? How big will the breach extension be? These questions have no answer yet. Nevertheless, policy makers are required to implement flood risk management plans over large areas and there is very little knowledge about how uncertainties in system behavior affect decisions. Furthermore, the overall system uncertainty is exacerbated by the expected climate change and socio-economic development.
This is what policy analysts refer to as a deeply uncertain problem: a problem where analysts struggle to provide a proper and unique model formulation to describe the interaction among the model’s variables, and have no or limited knowledge about the probability distribution of the uncertainties involved.
The current work introduces an exploratory model for the study of large-scale flood risk systems with the aim of developing long-term flood risk management strategies. The model will serve stakeholders and policy makers in exploring trade-offs among outcomes given different system configurations, and will foster discussion and learning about system performance in the face of deeply uncertain futures. To this end, the model provides a spatially distributed characterisation of flood risk by explicitly focusing on upstream - downstream interactions. The performance of each management strategy is assessed with respect to a heterogeneous set of performance indicators, belonging to both the economic and social domains. In particular, cost-benefit indicators, which are commonly used in flood risk management studies, will be used next to indicators related to equity and casualties.
The model will help answering three main research questions: (1) Given the huge system uncertainty, how can policy makers develop robust long-term flood risk management plans? (2) Would the adoption of a system approach change current flood risk management policies? (3) How to best analyze trade-offs emerging from conflicting stakeholder’s values and priorities?
Future research will focus on the application of the proposed modelling framework on two case studies: the lower branch of the Rhine River (from Bonn, Germany to the river Delta in The Netherlands) and the Po River (Italy).