PSI - Issue 78

Matjaž Dolšek et al. / Procedia Structural Integrity 78 (2026) 1569 – 1576

1572

for BORIS results computation. BORIS2 also plans the application or adaption of consequence models to estimate human impacts, such as displaced or injured population and fatalities (Foudi et al., 2015; Rossi et al., 2024), using methodologies like damage ratio thresholds comparable to seismic D3 damage scale (EMS-98, Grünthal, 1998). The main output of this step are: (i) a set of loss maps and summarizing tables for each considered return period, including economic losses, displaced people, injuries, and fatalities (intensity-based loss estimation); (ii) the loss curves resulting from the time-based risk assessment for the selected time window. These results enable the identification of risk-based hotspots, later used for prioritization and resource planning (Steps 3 and 4), and the hazard scenario selection performed in Step 2.

(a) (b) Fig. 1. Spatial distribution of grid points used for the exposure and seismic hazard models in the risk assessment of residential buildings in the Italy-Slovenia (above) and Austria-Slovenia (below) cross-border regions, as considered in the projects: (a) BORIS and (b) BORIS2. 3.2. Definition of earthquake, flood and compound hazards BORIS2 procedure allows the computation of user-defined seismic scenarios (e.g. historical events, pre-defined events with known source parameters) as well as the identification of a representative earthquake scenario at a location of interest (e.g. risk – based hotspot) consistent with a selected return period defined through a probabilistic approach. Once the target ground-motion intensity measure is selected, in terms of intensity level, hazard disaggregation (McGuire, 2004) is applied, which allows the id entification of a “control” earthquake— characterized by a representative magnitude, distance and residual variability — statistically consistent with the target return period. This control event is then used to generate deterministic ground-motion fields via ground motion prediction equations. Alternatively, multiple ground-motion fields can be generated for a given control earthquake scenario, as demonstrated by Babič et al. (2021). These fields, computed for bedrock conditions, are subsequently adjusted with site-specific amplification factors, if available, to reflect local soil conditions across the grid cells and at EMS units’ locations. The resulting ground-motion fields, which are associated with the selected return period, are used for loss estimation. Although the computed scenarios may not capture the full variability or epistemic uncertainty of real events, they may ensure a consistent and operationally meaningful basis for seismic scenario generation aimed at risk prevention and preparedness. The flood scenario generation in BORIS2 follows a probabilistic framework designed to create a comprehensive catalogue of flood events, from which scenarios corresponding to specific return periods can be extracted. Starting from existing flood hazard maps and their hydrological data, an interpolation process of the total volume of water for each original flood map generates a dense set of hazard layers for return periods between 1 and 500 years (Polese et al., 2024). BORIS2 procedure for flood scenario generation is based on the methodology developed by Ghizzoni et al. (2012) where a multivariate statistical approach is applied to simulate thousands of spatially coherent flood events over a synthetic 3000-year time window. These events preserve both the spatial correlation and the statistical properties of the generated flood events. The event generation algorithm consists of multiple steps performing a best fitting transformation of the discharges associated to each hydrological unit into probabilities, followed by sampling through a multivariate Gaussian distribution, and finally reconversion into discharge and depth values. The resulting