The Hazard Module

The Role of the Hazard Module

The hazard module is the scientific engine of every catastrophe model. Its job is to characterise the physical risk — answering three fundamental questions about potential future events: where will they occur, how often, and how severe will they be? These three elements are closely related, and their modelling requires similar datasets drawn from decades of scientific observation.

Locations of Potential Future Events

The first task is defining the model domain — the geographic region over which hazard sources must be identified. For earthquakes in southern California, this means identifying all faults and seismic source zones with measurable impact on the building inventory. The rate at which ground motion attenuates (weakens) with distance typically determines the appropriate geographical extent.

Earthquake Source Identification

Identifying earthquake sources requires multiple data types. The most straightforward approach uses mapped fault locations — some faults, like the San Andreas in California, are visible on the surface. However, historical seismicity catalogs, fault trenching, subsurface sounding, and aerial photography are all used to detect faults that may not be surface-visible. A fault that has shown no earthquake activity within the Holocene period (roughly the last 10,000 years) is generally considered inactive.

Not all earthquakes happen on known faults. In such cases, modellers define area (polygonal) source zones and use the spatial distribution of past earthquakes within the zone to estimate future earthquake locations — smoothing the data to allow simulated events to occur anywhere within the zone, not only where they have historically occurred.

Paleoseismology

Paleoseismology — the study of prehistoric earthquakes — provides crucial data for estimating the frequency of large events whose recurrence interval exceeds the historical record. Evidence includes offsets in geological formations, evidence of rapid coastal uplift or subsidence, laterally offset stream valleys, and liquefaction artifacts such as sand boils. Paleoseismic studies have provided compelling evidence for estimating magnitudes and return periods of large earthquakes in the New Madrid Seismic Zone in the central United States.

Frequency of Occurrence

Determining the annual probability of occurrence is the most critical and uncertain aspect of the hazard module. It is critical because all damage and loss probabilities are directly linked to this value. The uncertainty arises from the scarcity of historical data for rare events and from the fact that the underlying physical mechanisms that control natural hazard occurrence are still only partially understood.

The Gutenberg-Richter Relationship

For earthquakes, the relationship between frequency and magnitude is commonly modelled using the Gutenberg-Richter magnitude distribution, combined with the concept of the characteristic earthquake — where a specific fault ruptures at fairly regular intervals producing events of similar magnitude. Key parameters include lower and upper bound magnitudes, the a-value (rate of earthquakes above a reference magnitude), and the b-value (the rate at which cumulative frequency decreases as magnitude increases).

Hurricane Frequency

For hurricanes, frequency of occurrence reflects regional climate conditions. Two critical conditions for tropical cyclone formation are a large expanse of warm ocean water (generally at least 80°F / 27°C) and the relative absence of vertical wind shear. The most active months are when ocean temperatures peak: August and September in the Northern Hemisphere, January and February in the Southern Hemisphere.

Severity and Local Intensity

Once source parameters are generated, the model propagates intensity across the affected area — estimating hazard severity at each location in the building inventory.

Earthquake Attenuation

Seismic waves propagate outward from a fault rupture, weakening as they travel. Attenuation equations mathematically describe the rate at which wave amplitude decreases with distance from the source. These equations are region-specific, incorporating source mechanism (thrust vs. strike-slip faulting), crustal properties, and local soil conditions. Soft soils can dramatically amplify ground motion — the 1985 Mexico City earthquake, originating 400km away, caused severe damage in the city because soft lake-bed soils trapped and amplified the arriving waves.

Hurricane Windfields

For hurricanes, once source parameters (central pressure, forward speed, radius of maximum winds) are established, the model simulates the storm's movement along a track and calculates local wind speeds. Key adjustments are made for storm asymmetry, filling rate (as the storm moves inland, central pressure rises and winds weaken), and surface terrain roughness. The rougher the terrain, the more quickly wind speeds dissipate. Wind duration also matters — a slower-moving storm of the same intensity will cause more damage than a fast-moving one, as demonstrated by Hurricane Georges (1998), which stalled over the Gulf Coast causing higher losses than its Category 2 intensity would typically suggest.