Hierarchical Modelling of Flood Risk for Engineering Decision Analysis

Abstract

Societies around the world are faced with flood risk, prompting authorities and decision makers to manage risk to protect population and assets. With climate change, urbanisation and population growth, flood risk changes constantly, requiring flood risk management strategies that are flexible and robust. Traditional risk management solutions, e.g. dike construction, are not particularly flexible, as they are difficult to adapt to changing risk. Conversely, the recent concept of integrated flood risk management, entailing a combination of several structural and non-structural risk management measures, allows identifying flexible and robust flood risk management strategies. Based on it, this thesis investigates hierarchical flood protection systems, which encompass two, or more, hierarchically integrated flood protection structures on different spatial scales (e.g. dikes, local flood barriers and dry-proofed buildings), which jointly reduce risk. Hierarchical flood protection systems offer several advantages compared to single-structure flood protection systems, since they can be precisiontailored to fit risk reduction requirements and allow for flexible adaption of the protection system to changing flood risk. In the presence of flood protection structures, flood development depends on the state of all protection structures in the system. As such, hazard is a function not only of rainfall and river discharge, but also of protection structures’ fragility. A methodology for flood risk analysis and decision analysis for hierarchical flood protection systems is proposed, which allows for joint consideration of hazard models and fragility models of protection structures. In the implementation of the flood risk analysis methodology several challenges are identified, two of which are addressed in the present thesis. First, design and optimisation of a hierarchical flood protection system generally entails decisions about structures at different spatial scales, which, in turn, may require risk assessment at different spatial resolutions and levels of detail. Consistent risk modelling at different spatial scales may therefore require up- and downscaling of data and models under due consideration of uncertainties and dependencies. In this thesis, a methodology is proposed for spatially disaggregating an aggregated building portfolio considering disaggregation uncertainty and spatial correlation. The methodology is applied to the disaggregation of portfolios of buildings in two communes in Switzerland. The relevance of disaggregation uncertainty to natural hazard risk assessment is illustrated with a simple flood risk assessment example. A second challenge - fragility and vulnerability modelling of all protection structures in the hierarchical flood protection system - is identified. To optimise the design of protection structures, fragility and vulnerability models must allow for consideration of decision alternatives. While such vulnerability models are available for large protection structures (e.g. dikes), engineering vulnerability models that allow considering the impact of flood proofing measures on residential building vulnerability seem to be lacking. Thus, a flood vulnerability modelling approach for residential buildings is proposed, which allows for detailed building and hazard characterisation and models damages though explicit consideration of damage processes. The modelling approach allows for describing the impact of flood proofing measures on building vulnerability and can be utilised as a basis for decision analysis. The concept and usefulness of hierarchical flood protection systems, as well as the implementation of the flood risk analysis methodology and the vulnerability modelling approach are illustrated with an example application. In summary, the present thesis provides a characterisation of hierarchical flood protection systems as well as several methodologies to model such systems. It aims at increasing understanding of hierarchical flood protection systems and provides modelling approaches to facilitate further research and the implementation of hierarchical flood protection systems in practice.