Comparing built-up area datasets to assess urban exposure to coastal hazards in Europe.

Information on urban land use, beyond the urban-rural dichotomy, can improve the assessment of potential impacts of coastal hazards by refining estimates of damages and supporting adaptation planning. However, the lack of a consistent definition of "urban" in previous studies has led to exposure estimates that vary considerably. Here, we explore the sensitivity of exposed population and built-up area in four settlement types, defined by four different built-up area datasets. We find large differences in the exposed population of up to 65% (127 million people) in the "Urban" class. The exposure estimates are highly sensitive to the density thresholds used to distinguish the settlement types, with a difference in exposed urban population of up to 53.5 million people when the threshold varies by 10%. We attribute the high sensitivity of the exposure estimates to the varying definitions of built-up area of the underlying datasets. We argue that the definition of urban land is crucial for coastal impact assessments and make recommendations for the use of the analyzed datasets.