Addressing data limitations and uncertainties in broad-scale coastal flood-risk assessments
Coastal flooding constitutes a major risk to all low-lying coastal areas around the world. This risk is expected to increase during the 21st century with rising sea-levels and future societal development. Broad-scale coastal flood risk assessments are essential for identifying regions most at risk and evaluating the effectiveness of coastal adaptation responses in reducing future coastal impacts. Despite recent advances in coastal flood risk modelling research, there are a number of methodological and data related constraints and limitations inherent in broad-scale studies that affect the accuracy of assessment findings. Understanding and communicating these uncertainties is necessary for effectively supporting decision-makers in developing long-term robust and flexible adaptation plans. However, most uncertainties involved in broad-scale assessments are not fully quantified and their relative importance often remain unexplored. This thesis contributes to improving our understanding of data uncertainties and addresses data limitations in broad-scale coastal flood risk assessments. In particular, this thesis (1) addresses data availability, consistency and reproducibility constraints, (2) extends existing data models and increase the level of detail of assessments and (3) explores and quantifies data uncertainties in broad-scale coastal flood risk studies. For this purpose, Chapter 1 summarizes the main data limitations and uncertainties inherent in each coastal flood risk component (coastal hazard, exposure and vulnerability) and its implications for broad-scale coastal flood risk assessments. Chapter 2 assesses sea-level rise related coastal flood impacts for Emilia-Romagna (Italy) using the Dynamic Interactive Vulnerability Assessment (DIVA) modelling framework and investigates the sensitivity of model results to four uncertainty dimensions, namely (1) elevation, (2) population, (3) vertical land movement, (4) scale and resolution of assessment. Results show that by the end of the century coastal flood impacts are most sensitive to variations in elevation input data, followed by vertical land movement data and population data. The choice of one digital elevation model over another can lead up to 45% differences in the total extent of the coastal flood plain. Further, the inclusion of human-induced subsidence rates in the input data increases the relative sea-level rise on average by 60cm in 2100, resulting in coastal flood impacts that are up to 25% higher, highlighting that the non-consideration of human-induced subsidence in broad-scale studies underestimates coastal flood impacts. Chapter 3 describes the development of the open-access, spatially-explicit Mediterranean Coastal Database (MCD) that contains consistent information in terms of resolution, quality, accuracy and format of around 160 parameters on characteristics of the natural and socio-economic coastal subsystems for the entire region. The MCD, as well as the code for all data processing steps, is publicly available in an online repository. Chapter 4 illustrates the development of a new set of spatially-explicit projections of urban extent for ten countries in the Mediterranean, with a high spatial (100m) and temporal resolution (5-year time steps). These future urban projections indicate that accounting for the spatial patterns of urban development can lead to significant differences in the assessment of future coastal urban exposure. Depending on the urban development scenario chosen, the exposure of certain coastal regions can vary by up to 104 percent until 2100. The urban extent projections spanning from 2025 to 2100 and the python code to set up the urban change model are available from a public repository. Finally, Chapter 5 summarizes the main findings and lessons learned from this thesis and highlights some key challenges related to data that require further research targeting.