Fragility Analysis Based on Damaged Bridges during the 2021 Flood in Germany

Featured Application The existing literature on bridge fragility curves for floods mainly uses analytical approaches. However, it is crucial to validate these models and to identify failure trends and patterns to detect vulnerabilities. Therefore, fragility curves obtained using data from actual collapses can be employed in CAT (catastrophe) models. Indeed, a gateway to faster recovery from bridge failures can be achieved by transferring the financial risk to insurance providers. Fragility curves allow the association of the hazard intensity to several damage levels, thus enabling the use of damage-loss equations.Abstract Floods trigger the majority of expenses caused by natural disasters and are also responsible for more than half of bridge collapses. In this study, empirical fragility curves were generated by referring to actual failures that occurred in the 2021 flood in Germany. To achieve this, a calibrated hydraulic model of the event was used. Data were collected through surveys, damage reports and condition ratings from bridge owners. The database comprises 250 bridges. The analysis revealed recurrent failure mechanisms belonging to two main categories: those induced by scour and those caused by hydraulic forcing. The severity of the damage was primarily dependent on the bridge typology and, subsequently, on the deck's weight. The analysis allowed us to draw conclusions regarding the robustness of certain bridge typologies compared to others for a given failure mechanism. The likelihood of occurrence of the triggering mechanism was also highlighted as a factor to consider alongside the damage probability. This study sheds light on existing vulnerabilities of bridges to river floods, discussing specific areas in which literature data are contradictory. The paper also strengthens the call for a shift towards a probabilistic approach for estimating hydraulic force in bridge design and assessment.