PSI - Issue 78

Yazdan Almasi et al. / Procedia Structural Integrity 78 (2026) 433–440

435

distances due to buoyant uplift (Araki et al., 2017; Fehr et al., 2005). Although the overall industrial concentration was lower than in Japan, tank flotation and sloshing failures were reported. As noted in post-disaster assessments, floating roof tanks either overturned or were displaced due to hydrostatic uplift forces and loss of friction at the base slab (Mebarki et al., 2016; Nishino et al., 2024; Vitale, Ricci, et al., 2024). Tsunami waves also introduced debris impacts and foundation scouring (Basco & Salzano, 2017; Mebarki et al., 2016), which compromised equipment stability and increased the likelihood of flammable or toxic substance release (Basco & Salzano, 2017; Vitale, Ricci, et al., 2024). The most critical effects observed include tank uplift and displacement due to loss of anchorage and buoyancy forces (Mebarki et al., 2016; Vitale, Ricci, et al., 2024), shell buckling and foundation failure under combined cyclic and hydrostatic loads (Basco & Salzano, 2017; Mebarki et al., 2014, 2016; Vitale, Ricci, et al., 2024), process disruption and toxic releases following pipe disconnections or mechanical damage (Basco & Salzano, 2017), ignition and fires, particularly in oil and gas facilities due to compromised containment systems (Basco & Salzano, 2017; Nishino et al., 2024; Vitale, Ricci, et al., 2024). These cascading effects underscore the vulnerability of industrial equipment to both sequential and concurrent loading conditions. Assessing Na-Tech risks requires quantifying both seismic and tsunami hazards with appropriate intensity measures and spatial resolution. For seismic hazard, the European Seismic Hazard Model 2020 (ESHM20) developed by Pitilakis et al. (2024) offers a detailed probabilistic framework. It provides ground-shaking parameters (PGA, Spectral accelerations (Sa)) at multiple return periods and is consistent with national zoning criteria. ESHM20 also includes logic-tree approaches to account for epistemic uncertainty and is applicable across the Euro-Mediterranean region, including areas with coastal industrial concentration. Importantly, ESHM20 facilitates hazard disaggregation, which allows analysts to identify dominant magnitude-distance combinations affecting critical infrastructure and apply these in fragility models for industrial components. For tsunami hazard, hazard curves are typically derived using offshore earthquake scenarios combined with bathymetric and topographic data to simulate wave generation, propagation, and inundation (Rahimi et al., 2025; Vitale, Ricci, et al., 2024). Deterministic and probabilistic tsunami hazard assessments (PTHA) are employed depending on data availability. Methods such as those proposed by De Risi and Goda (2016), Goda et al. (2014), and Park and Cox (2016) integrate probabilistic seismic source models with numerical wave simulations to produce spatially-resolved tsunami intensity measures, such as inundation depth, momentum flux, and flow velocity. Recent developments in conditional hazard modeling enable the estimation of tsunami intensity measures, such as inundation depth or wave height, based on prior seismic event characteristics. This approach allows more realistic multi-hazard simulations in Na-Tech risk assessments, particularly by linking shared seismic source models to both ground shaking and tsunami inundation scenarios (De Risi & Goda, 2016; Goda & De Risi, 2023; Vitale, Baltzopoulos, et al., 2024). The combined use of seismic hazard models (e.g., ESHM20) and tsunami hazard modeling enables researchers to capture cascading effects. Several studies have introduced conditional hazard models, particularly through frameworks such as Probabilistic Seismic and Tsunami Hazard Analysis (PSTHA), in which tsunami hazard is evaluated given a preceding seismic event, thereby reflecting the true sequential nature of the compound hazard. This integrated approach enables a more realistic assessment of earthquake–tsunami cascades by conditioning tsunami simulations on the stochastic characteristics of the triggering earthquake (De Risi & Goda, 2016; Goda & De Risi, 2023; Park et al., 2017; Vitale, Baltzopoulos, et al., 2024). Hazard quantification has direct implications for structural reliability, performance assessment, and design criteria in industrial facilities. Pitilakis et al. (2024) emphasize that seismic hazard disaggregation and spatial correlation are particularly relevant for infrastructure systems with geographically distributed components, such as industrial clusters, due to the limitations of coarse-resolution hazard models like ESHM20 in capturing site-specific variability. Accordingly, the seismic demand placed on industrial structures must be evaluated not only in terms of ground motion intensity but also by accounting for local soil conditions, potential site amplification due to geology and topography, and the spatial resolution of the hazard models employed. A growing number of studies have sought to formalize multi-hazard definitions and methodologies that jointly characterize earthquake and tsunami risks, demonstrating the importance of coordinated hazard modeling and the limitations of assessing each hazard in isolation (De Risi & Goda, 2016; Goda & De Risi, 2023; Park et al., 2017; Rahimi et al., 2025; Vitale, Baltzopoulos, et al., 2024). However, current industrial fragility models often fail to capture time dependent degradation following an initial earthquake, making them insufficient for sequential risk scenarios without modification. Despite methodological advances, full integration of sequential damage accumulation and time dependent performance degradation into risk models for industrial facilities remains an open challenge.