PSI - Issue 84

Alessandro Scala et al. / Procedia Structural Integrity 84 (2026) 497–504

500

Figure 3 shows a selection of well-known bridge failure cases included in the database, representative of different natural triggering mechanisms and collapse scenarios. These examples include failures caused by debris flows, river flooding, slow landslide–structure interaction, and overtopping phenomena, and are intended to provide a qualitative reference framework for the types of events considered in the study. For each structure, where available, information was collected regarding the date and location of the event, the triggering cause, the structural typology, the construction material, and the observed level of damage. An objective inclusion criterion was adopted, namely the occurrence of a total or partial traffic disruption following the event, which was assumed to indicate a structurally and functionally significant level of damage. While this criterion ensures consistency across cases, it also implies an inherent bias toward events that attracted media or institutional attention, particularly for less severe damage scenarios. The spatial distribution of the full dataset is shown in Figure 4a, which provides an overview of bridge failures and severe damage events across the Italian territory, as derived from Scala et al. (2025). The distribution reflects the combined influence of geomorphological setting, infrastructure density, and exposure to hydrogeological hazards, and confirms the nationwide relevance of the collected cases. Damage severity was classified into three discrete levels, shown in Figure 4b. Damage Level 1 (DL1) includes cases characterized by structural or functional distress that compromised serviceability without leading to collapse. Damage Level 2 (DL2) refers to partial collapses, involving the failure of one or more structural components or limited portions of the structure. Damage Level 3 (DL3) corresponds to total collapse of the main structural system.

a)

b)

Fig. 4. (a) Geographic distribution of bridge failures and severe damage events in Italy (from Scala et al., 2025); (b) Frequency distribution of damage levels in the full database distinguishing the causes triggering the distress.

At database scale, Damage Level 1 (DL1) accounts for approximately 20% of the recorded cases. This relatively limited proportion should be interpreted considering the origin of the database, which is largely based on events reported by newspapers or national media. Minor damage scenarios are therefore less likely to be included unless associated with traffic disruption or explicit safety concerns. Within the DL1 category, the three predominant triggering causes, namely floods, landslides, and anthropogenic factors, are approximately evenly distributed. Damage Levels 2 (DL2) and 3 (DL3) represent comparable shares of the dataset, accounting for roughly 40% of the cases each. However, their internal cause distributions differ markedly. In the DL2 category, failures are predominantly associated with hydraulic causes, while landslide-related cases are relatively limited. Conversely, landslides become significantly more relevant in the DL3 category, where they contribute a substantial fraction of total collapses. Hydraulic-related failures remain the dominant cause across both DL2 and DL3 and exhibit a clear increase in relative importance with increasing damage severity. Although Figure 4b shows that anthropogenic causes represent the second most frequent triggering mechanism for bridge collapses overall, Scala et al. (2025) emphasized that, when climate change is considered as a boundary condition influencing structural vulnerability, hydraulic processes and landslide phenomena emerge as the most