NFPA Journal

NFPA reports annually on large-loss fires and explosions that occurred in the United States the previous year. These fires are defined as events that result in property damage of at least $10 million. In 2018, there were 36 large-loss fires, the highest number of such incidents since 2007 when there were 47 large-loss fires. The large-loss fires of 2018 resulted in more than $12.91 billion in direct property damage and losses, the highest figure since the $34.38 billion in losses experienced in 2001, which included losses associated with the attacks of 9/11. This makes 2018 the second-costliest year since the large-loss study started in 1975.

The three largest-loss fires in 2018 were all wildfires that occurred in wildland/urban interface (WUI) communities in California. The Camp Fire, the Woolsey Fire, and the Carr Fire were recognized as the second-, third-, and twelfth-costliest wildfires in US history, respectively. Three additional California wildfires—the Ranch Fire, the West Fire, and the Cranston Fire—were also classified as large-loss fires in this study. These six fires resulted in a combined property loss of $12.4 billion.

Firefighters work a blaze that occurred at a sawmill in Oregon. The fire resulted in $15 million in losses. ANDY NELSON/THE REGISTER GUARD VIA AP

Major wildfire incidents have been especially frequent over the last two years. Wildfires caused 94 percent and 96 percent of all property damage reported in the 2017 and 2018 large-loss studies, respectively. With $8.5 billion in losses, the Camp Fire was the third-largest large-loss incident in NFPA records, surpassed only by the 2001 terrorist attack on the World Trade Center and the 2017 October Fire Siege, a series of wildfires in California. Ignited on the same day as the Camp Fire, the Woolsey Fire spread throughout Southern California, resulting in over $2.9 billion in losses.

According to “Fire Loss in the United States During 2018” (September/October NFPA Journal), US fire departments responded to an estimated 1,318,500 fires that resulted in estimated losses of $25.6 billion. Many of these fires were small or resulted in little or no reported property damage. Although the 36 large-loss fires accounted for 0.003 percent of the estimated number of fires in 2018, they accounted for 50.4 percent of the total estimated dollar loss. In human terms, these 36 fires accounted for 102 deaths, 97 of which occurred in the wildland incidents. There were also at least 66 injuries. There were 14 more large-loss fires in 2018 than in 2017—eight structure fires and six non-structure fires—with an increase of more than $340 million in losses.

Where the fires occurred and how

Of the 36 large-loss fires in 2018, 28 involved structures and resulted in a total property loss of $529 million, or 4 percent of the combined losses for all large-loss fires. The other non-structural fires included six WUI fires and two vehicle fires that resulted in combined losses of $12.4 billion, or 96 percent of the losses in all the large-loss fires. The vehicles involved were a cargo ship and an aircraft.

Of the 28 structure fires, eight occurred in residential properties and accounted for $136.3 million in losses. Of those eight, seven occurred in apartment buildings, with the number of units ranging from 23 to 162; the number of units was not reported in one fire. One residential fire involved a single-family home.

A firefighter douses hot spots at the site of a large fire at a manufacturing facility in Oregon. CHRIS PIETSCH/THE REGISTER-GUARD VIA AP

Five fires occurred in manufacturing plants, causing a combined loss of $80.1 million. The facilities included two wood-product manufacturing plants, a saw mill, an electric products manufacturer, and a printing plant.

Four fires each occurred in public assembly, special, and storage properties. The public assembly property fires included two restaurants, a church, and a library, with a property loss of $114 million. The special properties included two motels under renovation, a boiler/mechanical room under construction, and a 180-unit apartment building under construction, with a total property loss of $96.2 million. The storage properties included two warehouses, a public works garage, and a lumber yard, with a combined total loss of $51.2 million.

Two fires occurred in stores and office properties: one in a greenhouse and the other in an automobile dealership, with a combined loss of $25 million.

One fire occurred on an industrial property—a hog farm—with a property loss of $20 million.

In 21 of the 36 large-loss fires last year, the cause and origin of the fires were undetermined, unknown, or were not reported. In several cases, the destruction was so extensive that investigators could not make a definitive cause determination or could not rule out several possibilities. Other fires are still under investigation or causes have not yet been reported.

The cause was reported for 12 of the 28 structure fires. Three of the fires were incendiary: two in occupied apartment buildings, and one in the apartment building that was under construction. Three fires were caused by hot work too close to combustibles: sweating pipes in an occupied apartment building, roof work on a church, and cutting close to a wall assembly in a storage building. Two fires were caused by careless disposal of smoking materials: one in a flower pot on the balcony of an apartment building, and the other in a greenhouse. One fire each was caused by a child playing with a lighter that ignited a mattress; a gas leak ignited by an unknown source at a construction site; an electrical malfunction in a digital technology manufacturing plant; and a candle located too close to curtains.

A fire involving a large single-family home in California resulted in more than $20 million in losses. Two firefighters were injured in the blaze. AP PHOTO/REED SAXON

The cause was reported for six of the eight non-structure fires. In the wildland fires, one fire each was caused by sparks from the rim of a blown-out trailer tire; an electrical malfunction; an electrical transmission line falling into dry brush; and a homeowner causing sparks while pounding a stake with a hammer. One wildland fire was deliberately set. The two vehicle fires were caused by hot work in the cargo area of a ship and a malfunction in an aircraft engine.

Operating status was reported for 21 of the 28 structure fires. In eight cases, the facility was open, operating, or had workers on-site; in 13 of the incidents, the properties were closed and/or unoccupied.

Seven of the 28 structure fires broke out between 11 p.m. and 7 a.m. and had a total direct property loss of $155.2 million. Two of the non-structure fires broke out between 11 p.m. and 7 a.m. with a total direct property loss of $8.4 billion.

Smoke detection and automatic
suppression equipment

Information about automatic fire or smoke detection equipment was reported for 24 of the 30 large-loss structure and vehicle fires. Of those 24 fires, 17 properties had detection equipment present, while seven properties had no automatic detection equipment installed. Ten of the 17 systems operated as designed, three systems did not operate, and in four cases the operation was not known. The three systems that did not operate were installed but had not yet been activated.

Information about automatic suppression equipment was reported for 24 of the 30 structure and vehicle fires. In those 24 fires, 11 properties had suppression systems present and 13 had no systems present. In three of the fires, the system operated, with one system extinguishing the fire—though the loss was high due to the cargo consisting of wind turbine components that were damaged—and the other two helping to control fire spread. In six cases, the systems did not operate: three because the sprinkler systems were installed but not yet activated due to construction; one because the system was not in the area of the fire; and two because the system failed for unreported reasons. The operation of the systems in the other two fires was not reported.

Complete information on the presence of both detection and suppression equipment was reported for 22 of the 30 structure fires. Nine structures had both detection and suppression equipment, seven had just detection equipment, five had neither system, and one had only suppression equipment.

What we can learn, where we get our data, and acknowledgements

Adhering to the fire protection principles reflected in NFPA codes and standards is essential if we are to reduce the occurrence of large-loss fires and explosions in the US. Proper construction, use of equipment, and procedures in chemical processes, storage, and housekeeping will make fires less likely to occur and help limit fire spread should a fire occur. Proper design, maintenance, and operation of fire protection systems and features can keep a fire that does occur from becoming a large-loss fire.

NFPA identifies potential large-loss incidents by reviewing national and local news media, as well as fire service publications. A clipping service reads all US daily newspapers and notifies NFPA’s Applied Research Division of major large-loss fires. NFPA’s annual survey of the US fire experience is an additional data source, although not the primary one. Web searches have proven useful in several cases where fire department and government reports have been released and published.

Once a fire has been identified, NFPA requests information about it from the fire department or jurisdictional agency. We also contact federal agencies that have participated in investigations, as well as state fire marshals’ offices and military sources. The diversity and redundancy of these data sources enables NFPA to collect the most complete data available on large-loss fires. This report includes only fire incidents for which NFPA has official dollar-loss estimates; other fires with large losses may have occurred, but are not included here because no official information has been reported to NFPA.

Due to a lack of a confirmed dollar loss, several 2018 fires that may have resulted in property losses greater than $10 million were not included in this report. These fires include a coke plant in Pennsylvania; a grain elevator and refinery in Wisconsin; a natural gas incident involving over-pressurization and explosions in Massachusetts; a warehouse in Washington; a metal products plant in Michigan; a cotton storage facility in Oklahoma; and a power generation plant in New Mexico

NFPA would like to thank the US fire service for its contributions of data, without which this report would not be possible. In some cases, the fire department, forestry officials, or government officials were unable to contribute complete details to NFPA because legal action is pending or ongoing, the incident was of a sensitive nature, or the size of the situation was overwhelming and reports had not yet been released. The authors also wish to thank Nancy Schwartz and the staff of the NFPA Applied Research group for providing support for this study. 

STEPHEN G. BADGER is a fire data assistant in NFPA’s Research Group and is a retired firefighter from the Quincy, Massachusetts, Fire Department.

MATTHEW FOLEY is a research associate in NFPA's Applied Research Group. Top photograph: Getty Images