At 5:12 a.m. on April 18, 1906, San Francisco residents awoke to the shaking of a massive M7.8 earthquake. Structures throughout the city were destroyed, and for at least four days, 50–60 fires burned and devastated the shattered city. The damage done by the conflagration fires that followed the 1906 San Francisco earthquake significantly exceeded the shake losses.
Today, insurance companies and catastrophe modeling firms look back at the 1906 San Francisco Earthquake and Fire as a turning point for how the industry assesses and mitigates risk from fire following earthquake. In the United States, standard home and renters insurance policies cover fire, resulting in insurance companies taking on a significant burden of risk.
Modeling this risk using an abundance of scientific and engineering research enhances industry preparedness to meeting the needs of both clients and society in the event of a catastrophe. To get the best understanding of the risk you’re covering, AIR is updating the fire following module in the 2017 AIR U.S. earthquake model, including a substantially increased resolution, a cutting-edge approach to fire ignition and spread modeling, and a comprehensively updated fire suppression component.
Resolution Matters
The risk associated with fire following earthquakes is intrinsically linked to the characteristics of the individual buildings in which fires might ignite and/or spread. To model the fire following sub-peril at high resolution, AIR developed a method to provide a detailed view of how buildings are arranged within city blocks. AIR scientists analyzed hundreds of city blocks and identified 20 representative block types (Figure 1).
This set of characteristic block configurations reflects building occupancies, heights and sizes of structures, and layouts. Land use and building density data are combined with the characteristic block parameters to produce a modeled distribution of 170 characteristic blocks of buildings or open space for every 1-km grid cell in the country. This approach to spatial distribution modeling at high resolution enables AIR to employ fire ignition and fire spread models at similarly high resolutions for more accurate risk estimation.
One of the most significant enhancements in the model is the ignition rate function. Post-earthquake ignitions are modeled using high resolution data from several U.S. earthquakes (such as 1989 Loma Prieta and 1994 Northridge earthquakes) to more accurately simulate the frequency and location of ignitions following a seismic event.
Fire spread modeling has evolved significantly from the traditionally-used empirically based equations, which fail to capture the full range of potential outcomes. To model fire following countrywide, AIR scientists have developed a cellular automata model. This is a physics-based modeling approach that captures fire spread on a 3-meter grid and is less computationally and data intensive than physics-based models, and is far more realistic than traditional approaches.
The new module uses 100-meter resolution land use data to identify the location of alternative water sources (i.e., fire stations and Auxiliary Water Suppression Systems) and accounts for the possibility of these resources being used for fire suppression. AIR used ISO® and Claritas data for the frequency and location of fire stations in the U.S. Additionally, the specially designed water suppression system for fire following earthquake in San Francisco—the Auxiliary Water Supply System (AWSS)—is also explicitly incorporated in the fire suppression model. With these updates, the 2017 fire following module for the upcoming AIR U.S. earthquake model provides a more detailed and accurate view of the risk associated with all identified impacts from fire following earthquake.
