Knowledge on flood risk, typically originating from data on past hazard, is a key input for developing disaster risk reduction strategies. With the observed increase in frequency and magnitude of flood hazard globally, knowledge on characterizing flood and predicting flood damage is especially important for data-scarce regions, which in many cases are regions with limited capacity to cope with disaster.

The objective of this thesis is to develop and test new methods, with reduced data requirement, which can improve characterizing floods and predicting flood damage in data-scarce areas. This overall objective was further divided into three sub-objectives: i) to review existing methods for assessing physical vulnerability to floods and conceptualize a new method for flood damage prediction, ii) to develop and test a hydrodynamic modelling approach to reconstruct a past flood scenario in a data-scare location, and iii) to test the applicability and performance of the new flood damage model and compare its performance to existing methods. Two study regions were selected from Nigeria to test the applicability of the developed methods. Both regions are located in the central part of the country, where small and large rivers are present and fluvial flooding is common.

The thesis has contributed to knowledge about the physical vulnerability of buildings to floods in typical data-scarce regions particularly in i) developing a new flood damage prediction method tailored for typical data-scarce areas with a transferable and up-datable framework, and ii) developing a method for increasing sample size of data (flood depth and duration) using hydrodynamic modelling to support physical vulnerability assessments. In addition, the thesis extensively discusses common challenges associated with flood risk assessment in data-scarce areas and suggests recommendations for future studies. The thesis presents one of the first findings for typical buildings in many African countries (sandcrete block and clay buildings) in terms of damage grades classification, identification of main damage drivers, and a model for predicting probable damage. The findings will support decision makers in data-scarce regions to better identify vulnerable buildings and plan for effective risk reduction and mitigation strategies.