The Camp Fire of November 2018 was the most deadly and costly wildfire in the recorded history of California wildfires. Extremes of weather, fuel, and complex topography came together to create a worst-case scenario. Following the fire, scientists from the Insurance Institute for Business & Home Safety (IBHS) and several IBHS member companies conducted a post-event investigation to examine the factors that contributed to this destructive fire.

The following report summarizes the results of the investigation and offers insights on how communities could be better prepared for future wildfire events.

Key findings include:

• Evaluating wildfire risk should include use of improved fuel models and a better way to account for high-wind events, such as those that drove the Camp Fire. Because there was a history of fire around the city of Paradise, the community had taken the time to develop and practice an evacuation plan. Despite this, the fuel, weather, and ignition conditions on November 8, 2018 led to a disastrous fire.
• Defensible space was found to be an important mitigation tactic to protect against all three ignition exposures—embers, direct flame, and radiant heat. While most hardening strategies can protect against one or two of the ignition mechanisms, the ones that reduce the thermal exposure to the building are the most reliable. However, defensible space did not guarantee survival and quantifying the degree of associated risk reduction remains a challenge.
• Previous testing at the IBHS Research Center has accurately reproduced several of the ignition scenarios observed during the post-event investigation. The ability to control for specific variables in a research environment is vital to understanding what combination of mitigation actions are effective in reducing the vulnerability of homes and communities to wildfire.
• Firefighter intervention has been shown to be a key factor in preventing loss of houses. This intervention usually occurs at, or shortly after, the time of ignition and before the structure becomes fully involved. However, this type of roving structure protection is unlikely in all scenarios. In most instances, the need for this intervention could have been avoided through proper mitigation strategies.
• IBHS found that even when using machine learning techniques, the simple building attribute information offered little predictive value. IBHS built a classifying model using post-event data to determine how simple building attributes (such as vegetative clearance, roof and siding material, window type, and local topography) affected survivability and whether they could provide some predictive capability. The results showed that survivability is far more complex, with many different factors combining to determine whether or not a structure survived.