Essential vocabulary for the catastrophe modelling professional. Search or browse alphabetically.
A
Annual Average Loss (AAL)
The expected loss from catastrophe events in any given year, averaged over many years of simulation. Calculated as the area under the exceedance probability curve. A key metric for pricing and reserving.
Annual Exceedance Probability (AEP)
The probability that the total losses from all events in a given year will exceed a specified threshold. Accounts for multiple events occurring in the same year.
Attritional Loss
Routine, everyday losses (e.g. small property claims) that insurers can reliably predict from historical data. Contrasted with catastrophe losses, which are rare and correlated.
C
Catastrophe (Cat)
A single event — natural or man-made — that causes insured losses above a defined threshold (commonly USD 25m). Catastrophes are characterised by low frequency, high severity, and spatial correlation.
Cat Model
A probabilistic simulation tool used by insurers and reinsurers to estimate the distribution of losses from natural or man-made catastrophes. Consists of hazard, exposure, vulnerability, and financial modules.
Correlation (Spatial)
The tendency for many policies to incur losses simultaneously from a single event. High spatial correlation — as in earthquakes — is a key driver of catastrophe accumulations.
E
Exceedance Probability (EP) Curve
A graph showing the probability that losses will exceed various thresholds. Can be expressed on an annual basis (AEP) or per-event basis (OEP). The primary output of a cat model.
Event Set / Stochastic Catalog
A collection of synthetic simulated events generated by the hazard module, representing the full range of possible events weighted by their probability of occurrence.
Exposure
The portfolio of insured assets potentially at risk from a catastrophe event. Described by location, value, occupancy type, construction type, and age of building. Data quality is critical to model accuracy.
G
Ground-Up Loss
The total physical loss to a structure from a catastrophe event, before the application of any insurance policy terms (deductibles, limits, etc.).
Ground Motion (Earthquake)
The shaking of the Earth's surface caused by seismic waves from an earthquake. Characterised by metrics such as Peak Ground Acceleration (PGA) or Spectral Acceleration (SA).
H
Hazard
The physical peril — the natural or man-made phenomenon that can cause damage. In cat modelling, the hazard module simulates the intensity and spatial footprint of events across the stochastic catalog.
Hurricane / Tropical Cyclone
An intense, rotating storm system originating over warm tropical oceans. Called "hurricane" in the Atlantic, "typhoon" in the Pacific, and "tropical cyclone" in the Indian Ocean. A major peril for coastal insurers.
I
Insured Loss
The portion of total physical losses that is covered under insurance or reinsurance policies, after applying policy terms such as deductibles, limits, and sub-limits.
Intensity Measure
A metric that characterises the severity of a hazard at a specific location. Examples: wind speed (m/s) for hurricane, Peak Ground Acceleration (g) for earthquake, inundation depth (m) for flood.
L
Loss Cost
The expected annual loss expressed as a percentage of the total insured value. Used in premium rating. A loss cost of 0.5% on a USD 10m property implies an expected annual loss of USD 50,000.
Liquefaction
A process where saturated soil loses strength during earthquake shaking and behaves like a liquid. Can cause severe structural damage even with moderate ground motion levels.
M
Mean Damage Ratio (MDR)
The expected ratio of repair cost to replacement value for a given building type at a given intensity level. A key output of the vulnerability module.
Model Uncertainty
Uncertainty arising from the assumptions, simplifications, and limitations of the cat model itself, as distinct from natural variability in hazard or data uncertainty in exposure inputs.
O
Occurrence Exceedance Probability (OEP)
The probability that the single largest event in a year will exceed a given loss level. Used when the concern is a single catastrophic event rather than aggregate annual loss.
P
Peril
A specific type of natural or man-made hazard. Examples: earthquake, hurricane, flood, wildfire, terror. Cat models are peril-specific — a windstorm model uses different science than an earthquake model.
Probable Maximum Loss (PML)
The maximum loss expected with a given probability (or return period). Typically the 1-in-250 or 1-in-200 year loss. Used by reinsurers and regulators to assess capital adequacy.
Protection Gap
The difference between total economic losses from a catastrophe and the insured portion. A large protection gap means most loss falls on governments, charities, and individuals rather than insurers.
R
Reinsurance
Insurance purchased by an insurance company to limit its own exposure to catastrophe losses. Reinsurers take on a share of the risk in exchange for a portion of the premium.
Return Period
The average number of years between events of a given severity or greater. A 1-in-100-year event has a 1% probability of occurring in any given year. Does not mean it occurs exactly once per century.
S
Stochastic Modeling
An approach that uses large numbers of simulated random events — typically 10,000–100,000 years of synthetic history — to explore the full probability distribution of possible outcomes.
Secondary Perils
Hazards triggered by or related to a primary event. Tsunami from earthquake; storm surge from hurricane; fire-following-earthquake. Often modelled separately but can dominate total losses.
T
Tail Risk
The risk of extreme, low-probability losses in the tail of the loss distribution. Catastrophe modelling exists primarily to characterise and price tail risk, which is underrepresented in historical data.
Tsunami
A series of ocean waves caused by a large undersea earthquake, landslide, or volcanic eruption. Can travel across entire ocean basins and cause catastrophic coastal flooding far from the source.
V
Vulnerability Function
A mathematical relationship between hazard intensity and expected damage ratio for a specific type of asset. Derived from empirical post-event damage data, engineering analysis, or expert judgment.