Critical infrastructure (CI) is increasingly exposed to natural hazards that can lead to devastating impacts on society. Understanding the vulnerability of CI assets to hazards like floods, earthquakes, windstorms, and landslides is crucial for effective risk mitigation. Key to this is the availability of fragility and vulnerability curves that quantify potential physical damage across a range of hazard intensities.
Through a systematic literature review, we have compiled the most comprehensive publicly available database of over 1510 unique fragility and vulnerability curves for CI assets. This database covers energy, transport, water, waste, telecommunication, health, and education systems, enabling multi-hazard risk assessments for infrastructure damage. Our findings highlight the wealth of research on earthquake vulnerability, especially for the energy and transportation sectors. In contrast, telecommunication assets remain underrepresented, warranting increased attention.
The database standardizes the curves to facilitate easy adoption and comparison across regions. It also provides critical details on the hazard intensity measures, asset characteristics, damage states, and curve derivation methods. This resource can directly support infrastructure owners, operators, and risk modelers in making informed decisions to enhance the resilience of these vital systems.
Beyond the compilation, this review outlines key opportunities for further developing the database, such as incorporating emerging hazards, expanding infrastructure types, and addressing multi-hazard interactions. Ultimately, this work aims to empower the disaster risk community with a comprehensive, transparent, and accessible knowledge base for assessing and managing the physical vulnerability of critical infrastructure.
Data Collection Methodology
We conducted a thorough literature search to identify relevant studies reporting fragility and vulnerability curves for CI assets exposed to floods, earthquakes, windstorms, and landslides. Using a combination of keywords related to hazards, infrastructure, and vulnerability, we systematically screened over 31,250 records from Google Scholar. This process yielded 95 suitable literature sources, from which we extracted 803 unique curve sets, comprising both fragility and vulnerability curves.
To ensure consistency and usability, we standardized the curves by converting them to common units, expressing them in relative terms, and filling in missing information through best estimates. We also cataloged important details for each curve, such as the hazard intensity measure, infrastructure characteristics, damage states, and derivation methods. This metadata is provided in a publicly available database accessible through the Zenodo repository.
Vulnerability Factors
The compiled curves cover a wide range of CI assets, including power plants, substations, transmission lines, roads, railways, water treatment facilities, and educational/healthcare buildings. The database highlights several key factors influencing the vulnerability of these assets:
Hazard Intensity Measures: Earthquakes commonly use peak ground acceleration (PGA) or peak ground velocity (PGV), while flooding primarily relies on inundation depth. Landslides employ a variety of metrics, such as rainfall intensity, debris flow height, and permanent ground deformation.
Asset Characteristics: Parameters like construction material, structural design, elevation, and age can significantly affect the vulnerability. For example, anchored storage tanks exhibit lower seismic fragility compared to unanchored ones.
Damage States: Fragility curves typically provide probabilities of reaching discrete damage states, such as “slight,” “moderate,” and “complete.” Vulnerability curves, on the other hand, directly estimate the degree of physical damage.
Geographical Context: The database reflects a global coverage, with curves developed for specific regions as well as generalized for broader application. This allows for comparative assessments of vulnerability across different settings.
Damage Estimation Models
The database provides two main types of curves for estimating infrastructure damage:
Fragility Curves: These describe the probability of reaching or exceeding a certain damage state for a given hazard intensity. They are commonly used in the earthquake engineering community.
Vulnerability Curves: These directly relate the hazard intensity to the expected degree of physical damage, often expressed as a mean damage ratio. They are more prevalent in flood risk assessments.
To facilitate the use of fragility curves, the database includes supplementary information on damage-to-loss models. These models enable the translation of fragility curves into vulnerability functions by relating physical damage to monetary loss or other impact metrics.
Critical Infrastructure Components
Building Structures
The database includes vulnerability curves for various types of commercial and residential buildings, including schools, hospitals, and office facilities. These curves generally consider parameters like construction material, number of stories, and seismic design.
Commercial Facilities: Huizinga et al. (2017) provide depth-damage curves for the “commercial buildings” category, which encompasses schools and hospitals. Kok et al. (2005) present a generalized curve for companies and government buildings, including educational and healthcare institutions.
Residential Buildings: While the database does not focus on residential structures, the FEMA (2013, 2020, 2021a) and Milutinovic and Trendafiloski (2003) curves for the general building stock can be applicable to housing assets.
Transportation Systems
Transportation infrastructure, including roads, railways, and airports, is a critical component of CI. The database covers a range of vulnerability and fragility curves for these assets, highlighting their susceptibility to various natural hazards.
Road Networks: Huizinga (2007) and Huizinga et al. (2017) provide generalized depth-damage curves for roads, which are also applied to railways. Van Ginkel et al. (2021) developed more granular, asset-specific curves for different road types in Europe.
Rail Infrastructure: Kellermann et al. (2015, 2016) and Bubeck et al. (2019) present vulnerability functions for railways, considering both flood inundation and seismic ground shaking. Argyroudis and Kaynia (2014, 2015) offer fragility curves for railway embankments and cuts.
Airports: FEMA (2013, 2020, 2021a) developed vulnerability and fragility curves for airport runways, fuel facilities, and other structures, accounting for both flood and seismic hazards.
Hazard Exposure Analysis
Natural Disaster Scenarios
The database covers a range of natural hazards that can impact CI, including floods, earthquakes, windstorms, and landslides. These hazards differ in their characteristics and the mechanisms by which they can damage infrastructure assets.
Wind Storms: Vulnerability to wind loading is a significant concern for power plants, substations, and transmission towers/lines. The database includes fragility curves that consider factors like structural design, wind speed, and attack angle.
Heavy Precipitation: Flooding can severely impact road networks, water treatment facilities, and underground pipelines. The depth-damage curves in the database account for the inundation depth as the primary hazard intensity measure.
Geospatial Mapping
Assessing the vulnerability of CI requires detailed spatial information on asset locations and the surrounding hazard environment. The database supports this through several key features:
Asset Geo-Positioning: Many of the curves are developed for specific regions or countries, allowing for the integration of local infrastructure data and hazard mapping.
Hazard Mapping: The database covers a range of hazard types, enabling the combination of spatial hazard models (e.g., flood, earthquake, landslide) with the corresponding vulnerability curves.
Risk Quantification and Prioritization
Probabilistic Risk Assessment
The fragility and vulnerability curves in the database can be used as core inputs for probabilistic risk assessment frameworks. These enable the estimation of potential physical damage and associated economic losses under various hazard scenarios.
Fragility Curves: By providing probabilities of reaching damage states, fragility curves can be coupled with consequence models to quantify the likelihood and severity of infrastructure failures.
Vulnerability Curves: These directly translate hazard intensities into damage estimates, facilitating the calculation of direct economic losses to infrastructure assets.
Decision Support Tools
The comprehensive database can support the development of advanced decision support tools for infrastructure owners and risk managers. These tools can aid in prioritizing investments, optimizing mitigation strategies, and enhancing overall resilience.
Visualization Dashboards: Interactive visualizations of the vulnerability data can help stakeholders quickly identify high-risk assets and compare the performance of different infrastructure types.
Optimization Algorithms: The standardized curves can be integrated into optimization frameworks to identify cost-effective retrofitting or hardening measures for CI systems.
Through this extensive review and database, we have provided the disaster risk community with a robust resource for assessing the physical vulnerability of critical infrastructure. By synthesizing the state-of-the-art knowledge, we aim to empower infrastructure owners, operators, and risk modelers to make informed decisions that enhance the resilience of these vital systems in the face of natural hazards.
