From simulation to reality: Validation of fast flood modelling using fire station operational data.

Short-duration rainfall events that are increasingly extreme and the resulting pluvial flooding pose a significant threat to urban areas. Robust flood models, therefore, play a crucial role in urban stormwater management. However, the applicability of these models is strongly constrained by the accuracy and reliability of their predictions, making thorough validation essential. At a city level, validation can be difficult due to a lack of data and reference measurements. Consequently, alternative approaches that use real-event observations, such as fire station operational records, photographs, or videos, are becoming increasingly interesting. This study investigates the potential of fire station operational data, including photos, for validating the Dynamic CA-ffé fast flood model. Comparing the results with those of an existing coupled 1D/2D SWMM model provides additional context and enables an assessment of the suitability of such data for model validation. The results demonstrate that, while fire brigade records permit comparisons only for affected and documented buildings, they still offer a reliable indication of the overall flood pattern. Hit rates range from 34 to 89%, depending on the precipitation event and case study. Photographic material offers additional information on flood extent. However, several limitations were also identified, including a lack of information on non-reported or unaffected buildings. In conclusion, fire station operational data can be considered useful for coarsely validating urban flood models. However, definitive statements can only be derived if data collection is expanded to include non-reported flooded and non-affected buildings, as well as water level data and more photographic documentation.