New FSE wiki article: Fire cause (written by Claude Opus 5.5) |
Add SVG figures fse-fire-cause-v1.svg, fse-fire-cause-v2.svg, fse-fire-cause-v3.svg, fse-fire-cause-v4.svg (list-8 visual plan) |
||
| (One intermediate revision by one other user not shown) | |||
| Line 1: | Line 1: | ||
A '''fire cause''' is the combination of circumstances that leads to the ignition of an unwanted fire. It comprises a source of heat, a material that is ignited, and the human act, omission or equipment failure that brings the two together.<ref name="HallCote-3-11-12">Hall and Cote (2008), pp. 3-11–3-12.</ref> In the ''Fire Protection Handbook'' of the National Fire Protection Association (NFPA), Custer and Sheppard (2008) note that in most fires the origin and cause are expressed in four terms. These are the area of origin, the heat of ignition, the materials involved and the factors that brought them together.<ref name="CusterSheppard-3-31">Custer and Sheppard (2008), p. 3-31.</ref> Fire investigators determine the cause of individual fires, and fire data systems record it in coded form so that causes can be counted and ranked.<ref name="CusterSheppard-3-31" /><ref name="Ahrens-3-71">Ahrens et al. (2008), p. 3-71.</ref> | A '''fire cause''' is the combination of circumstances that leads to the ignition of an unwanted fire. It comprises a source of heat, a material that is ignited, and the human act, omission or equipment failure that brings the two together.<ref name="HallCote-3-11-12">Hall and Cote (2008), pp. 3-11–3-12.</ref><ref name="CusterSheppard-3-37">Custer and Sheppard (2008), p. 3-37.</ref> In the ''Fire Protection Handbook'' of the National Fire Protection Association (NFPA), Custer and Sheppard (2008) note that in most fires the origin and cause are expressed in four terms. These are the area of origin, the heat of ignition, the materials involved and the factors that brought them together.<ref name="CusterSheppard-3-31">Custer and Sheppard (2008), p. 3-31.</ref> Fire investigators determine the cause of individual fires, and fire data systems record it in coded form so that causes can be counted and ranked.<ref name="CusterSheppard-3-31" /><ref name="Ahrens-3-71">Ahrens et al. (2008), p. 3-71.</ref> | ||
There is no single list of causes, because a fire can be classified by its heat source, by the item first ignited, by the equipment involved or by the behaviour behind it.<ref name="Ahrens-3-71" /> In United States structure fires in 1999 to 2002, cooking equipment was the leading cause of fires and injuries, smoking materials of civilian deaths and intentional fires of direct property damage.<ref name="HallCote-3-14-15">Hall and Cote (2008), pp. 3-14–3-15.</ref> Knowing the cause of fires can lead to targeted inspection, public education and code changes, and fire engineers use cause data to choose fire scenarios and estimate how often fires start.<ref name="CusterSheppard-3-31" /><ref name="Hall2002b-5-147">Hall (2002b), p. 5-147.</ref> The statistics in this article describe the United States in the years around 2000 and come from sources published up to 2008. | There is no single list of causes, because a fire can be classified by its heat source, by the item first ignited, by the equipment involved or by the behaviour behind it.<ref name="Ahrens-3-71" /> In United States structure fires in 1999 to 2002, cooking equipment was the leading cause of fires and injuries, smoking materials of civilian deaths and intentional fires of direct property damage.<ref name="HallCote-3-14-15">Hall and Cote (2008), pp. 3-14–3-15.</ref> Knowing the cause of fires can lead to targeted inspection, public education and code changes, and fire engineers use cause data to choose fire scenarios and estimate how often fires start.<ref name="CusterSheppard-3-31" /><ref name="Hall2002b-5-147">Hall (2002b), p. 5-147.</ref> The statistics in this article describe the United States in the years around 2000 and come from sources published up to 2008. | ||
| Line 7: | Line 7: | ||
Hall and Cote (2008) build their account of fire prevention on three components. Every hostile fire requires an initial heat source, an initial fuel source and something to bring them together. That something nearly always has a human component, usually an immediate act or omission, or sometimes the delayed effect of an error in design or installation.<ref name="HallCote-3-11-12" /> Prevention can act on any of the three, and causes can be defined and ranked in terms of any of them.<ref name="HallCote-3-12">Hall and Cote (2008), p. 3-12.</ref><ref name="HallCote-3-14">Hall and Cote (2008), p. 3-14.</ref> | Hall and Cote (2008) build their account of fire prevention on three components. Every hostile fire requires an initial heat source, an initial fuel source and something to bring them together. That something nearly always has a human component, usually an immediate act or omission, or sometimes the delayed effect of an error in design or installation.<ref name="HallCote-3-11-12" /> Prevention can act on any of the three, and causes can be defined and ranked in terms of any of them.<ref name="HallCote-3-12">Hall and Cote (2008), p. 3-12.</ref><ref name="HallCote-3-14">Hall and Cote (2008), p. 3-14.</ref> | ||
[[File:fse-fire-cause-v1.svg|thumb|upright=1.5|alt=Diagram: heat source, fuel and a human or equipment factor combine into a fire cause, with the matching NFIRS data elements|A fire cause combines a heat source, a fuel and the human act, omission or equipment failure that brings them together; each maps onto NFIRS data elements.]] | |||
Investigators analyse the ignition sequence in the same terms. Custer and Sheppard (2008) list three items. The first is the size, nature and source of the ignition energy, and the second the physical and chemical fire properties of the material ignited. The third is the set of human or mechanical circumstances that brought energy and fuel together.<ref name="CusterSheppard-3-37">Custer and Sheppard (2008), p. 3-37.</ref> A candidate source must be a competent [[Ignition source|ignition source]], with enough energy and temperature to raise the fuel to its ignition temperature during the exposure time. An electric arc, for example, may be far hotter than the ignition temperature of a fuel but too brief to ignite it.<ref name="CusterSheppard-3-37" /> | Investigators analyse the ignition sequence in the same terms. Custer and Sheppard (2008) list three items. The first is the size, nature and source of the ignition energy, and the second the physical and chemical fire properties of the material ignited. The third is the set of human or mechanical circumstances that brought energy and fuel together.<ref name="CusterSheppard-3-37">Custer and Sheppard (2008), p. 3-37.</ref> A candidate source must be a competent [[Ignition source|ignition source]], with enough energy and temperature to raise the fuel to its ignition temperature during the exposure time. An electric arc, for example, may be far hotter than the ignition temperature of a fuel but too brief to ignite it.<ref name="CusterSheppard-3-37" /> | ||
| Line 14: | Line 15: | ||
== Determining the cause == | == Determining the cause == | ||
In the United States, the local fire department generally has the primary responsibility to document a fire and to make an initial determination of its origin and cause.<ref name="CusterSheppard-3-31" /> If the cause is easily determined and not criminal, the investigation usually ends there. When the cause cannot be determined, a death is involved or arson is suspected, additional help is generally called in.<ref name="CusterSheppard-3-38">Custer and Sheppard (2008), p. 3-38.</ref> When a fire may have been set deliberately, the state or provincial fire marshal's office is responsible for the investigation in much of the United States and in all Canadian provinces.<ref name="CusterSheppard-3-32">Custer and Sheppard (2008), p. 3-32.</ref> The methods of [[Fire investigation]] are described in NFPA 921, ''Guide for Fire and Explosion Investigations''.<ref name="FarrSawyer-1-83">Farr and Sawyer (2008), p. 1-83.</ref> | In the United States, the local fire department generally has the primary responsibility to document a fire and to make an initial determination of its origin and cause.<ref name="CusterSheppard-3-31" /> If the cause is easily determined and not criminal, the investigation usually ends there. When the cause cannot be determined, a death is involved or arson is suspected, additional help is generally called in.<ref name="CusterSheppard-3-38">Custer and Sheppard (2008), p. 3-38.</ref> When a fire may have been set deliberately, the state or provincial fire marshal's office is responsible for the investigation in much of the United States and in all Canadian provinces.<ref name="CusterSheppard-3-32">Custer and Sheppard (2008), p. 3-32.</ref> The methods of [[Fire investigation]] are described in NFPA 921, ''Guide for Fire and Explosion Investigations''.<ref name="FarrSawyer-1-83">Farr and Sawyer (2008), p. 1-83.</ref> | ||
[[File:fse-fire-cause-v3.svg|thumb|upright=1.5|alt=Flowchart of US fire cause determination from fire department to investigation and incident report|Steps in determining the cause of a fire in the United States.]] | |||
NFPA 921 was first issued in 1992 and was revised every three years up to 2004. Its 2004 edition included chapters on origin determination, fire cause determination, analysing the incident for cause and responsibility, and incendiary fires.<ref name="CusterSheppard-3-33">Custer and Sheppard (2008), p. 3-33.</ref> The guide also forms the basis for the performance criteria of NFPA 1033 for fire investigators, as described by Custer and Sheppard (2008).<ref name="CusterSheppard-3-33" /> | NFPA 921 was first issued in 1992 and was revised every three years up to 2004. Its 2004 edition included chapters on origin determination, fire cause determination, analysing the incident for cause and responsibility, and incendiary fires.<ref name="CusterSheppard-3-33">Custer and Sheppard (2008), p. 3-33.</ref> The guide also forms the basis for the performance criteria of NFPA 1033 for fire investigators, as described by Custer and Sheppard (2008).<ref name="CusterSheppard-3-33" /> | ||
| Line 22: | Line 24: | ||
== Cause categories in incident data == | == Cause categories in incident data == | ||
NFIRS records cause information in six data elements. These are a general cause category, the heat source, the equipment involved in ignition, the item first ignited, the type of material ignited, and the factors contributing to ignition.<ref name="Ahrens-3-70-71">Ahrens et al. (2008), pp. 3-70–3-71.</ref> Together these can define billions of distinct cause categories, so analysts use simpler groupings.<ref name="Ahrens-3-71" /> In Version 5.0 of NFIRS, the general cause field has the options intentional, unintentional, failure of equipment or heat source, act of nature, cause under investigation, other, and cause undetermined after investigation. | NFIRS records cause information in six data elements. These are a general cause category, the heat source, the equipment involved in ignition, the item first ignited, the type of material ignited, and the factors contributing to ignition.<ref name="Ahrens-3-70-71">Ahrens et al. (2008), pp. 3-70–3-71.</ref> Together these can define billions of distinct cause categories, so analysts use simpler groupings.<ref name="Ahrens-3-71" /> In Version 5.0 of NFIRS, the general cause field has the options intentional, unintentional, failure of equipment or heat source, act of nature, cause under investigation, other, and cause undetermined after investigation. Cause under investigation and cause undetermined after investigation are both treated as unknown in analysis.<ref name="AhrensStewart-3-50">Ahrens and Stewart (2008), p. 3-50.</ref><ref name="Ahrens-3-66">Ahrens et al. (2008), p. 3-66.</ref><ref name="Ahrens-3-71" /> | ||
Intentional fires need careful reading. Older versions of NFIRS used "incendiary" when a legal decision or physical evidence showed that a fire was deliberately set, and "suspicious" when circumstances only suggested it.<ref name="AhrensStewart-3-49" /> Analysts often grouped both as arson, although NFPA 921 does not consider "suspicious" a valid cause.<ref name="Ahrens-3-68">Ahrens et al. (2008), p. 3-68.</ref> Version 5.0 replaced both with "intentional", which does not imply an illegal act as strongly, and none of these terms corresponds exactly to arson.<ref name="AhrensStewart-3-49" /><ref name="Ahrens-3-68" /> An intentional fire set by a young child, for example, may not be arson in the legal sense.<ref name="AhrensStewart-3-49" /><ref name="Ahrens-3-72">Ahrens et al. (2008), p. 3-72.</ref> | Intentional fires need careful reading. Older versions of NFIRS used "incendiary" when a legal decision or physical evidence showed that a fire was deliberately set, and "suspicious" when circumstances only suggested it.<ref name="AhrensStewart-3-49" /> Analysts often grouped both as arson, although NFPA 921 does not consider "suspicious" a valid cause.<ref name="Ahrens-3-68">Ahrens et al. (2008), p. 3-68.</ref> Version 5.0 replaced both with "intentional", which does not imply an illegal act as strongly, and none of these terms corresponds exactly to arson.<ref name="AhrensStewart-3-49" /><ref name="Ahrens-3-68" /> An intentional fire set by a young child, for example, may not be arson in the legal sense.<ref name="AhrensStewart-3-49" /><ref name="Ahrens-3-72">Ahrens et al. (2008), p. 3-72.</ref> | ||
| Line 53: | Line 55: | ||
Each cause dominates a different measure of loss. Cooking equipment is the leading cause of structure fires and structure fire injuries, and it is involved in the majority of unreported home fires. Unattended cooking is the principal behavioural factor in these [[Cooking fire]]s.<ref name="HallCote-3-15">Hall and Cote (2008), p. 3-15.</ref> Smoking materials are the leading cause of civilian fire deaths, and most of these [[Smoking-material fire]]s begin with the ignition of upholstered furniture, mattresses or bedding.<ref name="HallCote-3-14" /> Intentional fires are the leading cause of structure fire property damage, accounting for nearly one dollar in every five lost.<ref name="HallCote-3-14-15" /> Electrical distribution equipment ranks no higher than fourth on any measure except property damage, where it ranks second.<ref name="HallCote-3-15" /> | Each cause dominates a different measure of loss. Cooking equipment is the leading cause of structure fires and structure fire injuries, and it is involved in the majority of unreported home fires. Unattended cooking is the principal behavioural factor in these [[Cooking fire]]s.<ref name="HallCote-3-15">Hall and Cote (2008), p. 3-15.</ref> Smoking materials are the leading cause of civilian fire deaths, and most of these [[Smoking-material fire]]s begin with the ignition of upholstered furniture, mattresses or bedding.<ref name="HallCote-3-14" /> Intentional fires are the leading cause of structure fire property damage, accounting for nearly one dollar in every five lost.<ref name="HallCote-3-14-15" /> Electrical distribution equipment ranks no higher than fourth on any measure except property damage, where it ranks second.<ref name="HallCote-3-15" /> | ||
[[File:fse-fire-cause-v2.svg|thumb|upright=1.5|alt=Grouped bar chart of seven US fire causes: shares of fires, deaths and property damage differ by cause|Share of US structure fires, civilian deaths and direct property damage by leading cause, 1999–2002 annual averages, excluding 11 September 2001.]] | |||
Severity per fire differs widely between causes. In US home structure fires in 1999 to 2002, smoking materials caused 3.4 civilian deaths per 100 fires, intentional fires and playing with a heat source 1.6, and cooking equipment 0.3.<ref name="Ahrens-3-74" /> Intentional fires, candle fires and electrical distribution or lighting fires had the largest average property loss per fire, about $21,000 to $21,500.<ref name="Ahrens-3-74" /><ref name="Ahrens-3-76">Ahrens et al. (2008), p. 3-76.</ref> Outside homes, intentional fires led in both the number of fires and property damage, with 18 percent of nonhome structure fires and 27 percent of their direct property damage.<ref name="HallCote-3-16">Hall and Cote (2008), p. 3-16.</ref> | Severity per fire differs widely between causes. In US home structure fires in 1999 to 2002, smoking materials caused 3.4 civilian deaths per 100 fires, intentional fires and playing with a heat source 1.6, and cooking equipment 0.3.<ref name="Ahrens-3-74" /> Intentional fires, candle fires and electrical distribution or lighting fires had the largest average property loss per fire, about $21,000 to $21,500.<ref name="Ahrens-3-74" /><ref name="Ahrens-3-76">Ahrens et al. (2008), p. 3-76.</ref> Outside homes, intentional fires led in both the number of fires and property damage, with 18 percent of nonhome structure fires and 27 percent of their direct property damage.<ref name="HallCote-3-16">Hall and Cote (2008), p. 3-16.</ref> | ||
| Line 61: | Line 64: | ||
Many fires are reported with the cause, or part of it, unknown. In nonconfined US home structure fires of 1999 to 2002, the item first ignited was undetermined or not reported in 21 percent of fires, but in 39 percent of the civilian deaths.<ref name="Ahrens-3-76" /> Severe fires tend to have larger shares of unknown data, because more of the evidence of origin is destroyed.<ref name="Ahrens-3-76" /> Since Version 5.0 of NFIRS, confined fires, such as cooking fires confined to their container, may be reported without causal data.<ref name="Ahrens-3-66" /> Their share of structure fires in NFIRS rose from 2.8 percent in 1999 to 34.2 percent in 2003, largely because more departments used Version 5.0.<ref name="Ahrens-3-76-77">Ahrens et al. (2008), pp. 3-76–3-77.</ref> | Many fires are reported with the cause, or part of it, unknown. In nonconfined US home structure fires of 1999 to 2002, the item first ignited was undetermined or not reported in 21 percent of fires, but in 39 percent of the civilian deaths.<ref name="Ahrens-3-76" /> Severe fires tend to have larger shares of unknown data, because more of the evidence of origin is destroyed.<ref name="Ahrens-3-76" /> Since Version 5.0 of NFIRS, confined fires, such as cooking fires confined to their container, may be reported without causal data.<ref name="Ahrens-3-66" /> Their share of structure fires in NFIRS rose from 2.8 percent in 1999 to 34.2 percent in 2003, largely because more departments used Version 5.0.<ref name="Ahrens-3-76-77">Ahrens et al. (2008), pp. 3-76–3-77.</ref> | ||
The usual remedy is to allocate the unknowns in proportion to the known cases. This assumes that the fires with unknown cause, if their cause were known, would show the same profile as those with known cause.<ref name="Ahrens-3-76" /> | The usual remedy is to allocate the unknowns in proportion to the known cases. This assumes that the fires with unknown cause, if their cause were known, would show the same profile as those with known cause.<ref name="Ahrens-3-76" /> As an illustration of proportional allocation for fire counts, the allocated share of cause category <math>c</math> is its share among fires with known cause:<ref name="Ahrens-3-76" /> | ||
<math display="block">p_c = \frac{n_c}{N - n_u}, \qquad \hat{N}_c = p_c \, N</math> | <math display="block">p_c = \frac{n_c}{N - n_u}, \qquad \hat{N}_c = p_c \, N</math> | ||
| Line 68: | Line 71: | ||
! Symbol !! Description !! Unit | ! Symbol !! Description !! Unit | ||
|- | |- | ||
| <math>N</math> || Total number of fires | | <math>N</math> || Total number of fires in the data set || fires | ||
|- | |- | ||
| <math>n_u</math> || Number with cause unknown, undetermined or not reported || fires | | <math>n_u</math> || Number with cause unknown, undetermined or not reported || fires | ||
| Line 80: | Line 83: | ||
Ahrens et al. (2008) give a hypothetical example of two neighbouring states with the same underlying cause profile.<ref name="Ahrens-3-76" /> State A reports 9 percent of home fires as intentional and 14 percent with unknown cause, while State B reports 6 percent intentional and 43 percent unknown. Before allocation State B seems to have fewer intentional fires. As an illustration of the relation, State A gives <math>p = 9/(100-14) = 0.105</math> and State B gives <math>p = 6/(100-43) = 0.105</math>, so both states have about 10.5 percent intentional fires. The same holds for the other categories, which confirms the authors' point that both states have the same problems to the same degree.<ref name="Ahrens-3-76" /> | Ahrens et al. (2008) give a hypothetical example of two neighbouring states with the same underlying cause profile.<ref name="Ahrens-3-76" /> State A reports 9 percent of home fires as intentional and 14 percent with unknown cause, while State B reports 6 percent intentional and 43 percent unknown. Before allocation State B seems to have fewer intentional fires. As an illustration of the relation, State A gives <math>p = 9/(100-14) = 0.105</math> and State B gives <math>p = 6/(100-43) = 0.105</math>, so both states have about 10.5 percent intentional fires. The same holds for the other categories, which confirms the authors' point that both states have the same problems to the same degree.<ref name="Ahrens-3-76" /> | ||
[[File:fse-fire-cause-v4.svg|thumb|upright=1.5|alt=Stacked bars for two states before and after allocating unknown causes; after allocation the cause shares are equal|Hypothetical cause profiles of home fires in two states before and after proportional allocation of unknown causes.]] | |||
No allocation method has been shown to be best, and some analysts argue that time of day, fire size or property type could be used to allocate unknowns differently.<ref name="Ahrens-3-76" /> Because cause profiles differ between large and small fires, an allocation ideally takes fire size into account, although no consensus on this had formed.<ref name="Ahrens-3-76" /> NFPA allocates the heat source code "other open flame or smoking material" in proportion to the specific codes. Without this secondary allocation, intentional fires would overtake smoking materials as the leading cause of home fire deaths.<ref name="Ahrens-3-77">Ahrens et al. (2008), p. 3-77.</ref> In national estimates, Hall and Ahrens (2002) describe the same allocation as built into the scaling ratios that project NFIRS data to the totals of the NFPA survey.<ref name="HallAhrens-5-67-68">Hall and Ahrens (2002), pp. 5-67–5-68.</ref> | No allocation method has been shown to be best, and some analysts argue that time of day, fire size or property type could be used to allocate unknowns differently.<ref name="Ahrens-3-76" /> Because cause profiles differ between large and small fires, an allocation ideally takes fire size into account, although no consensus on this had formed.<ref name="Ahrens-3-76" /> NFPA allocates the heat source code "other open flame or smoking material" in proportion to the specific codes. Without this secondary allocation, intentional fires would overtake smoking materials as the leading cause of home fire deaths.<ref name="Ahrens-3-77">Ahrens et al. (2008), p. 3-77.</ref> In national estimates, Hall and Ahrens (2002) describe the same allocation as built into the scaling ratios that project NFIRS data to the totals of the NFPA survey.<ref name="HallAhrens-5-67-68">Hall and Ahrens (2002), pp. 5-67–5-68.</ref> | ||
Latest revision as of 18:08, 14 November 2026
A fire cause is the combination of circumstances that leads to the ignition of an unwanted fire. It comprises a source of heat, a material that is ignited, and the human act, omission or equipment failure that brings the two together.[1][2] In the Fire Protection Handbook of the National Fire Protection Association (NFPA), Custer and Sheppard (2008) note that in most fires the origin and cause are expressed in four terms. These are the area of origin, the heat of ignition, the materials involved and the factors that brought them together.[3] Fire investigators determine the cause of individual fires, and fire data systems record it in coded form so that causes can be counted and ranked.[3][4]
There is no single list of causes, because a fire can be classified by its heat source, by the item first ignited, by the equipment involved or by the behaviour behind it.[4] In United States structure fires in 1999 to 2002, cooking equipment was the leading cause of fires and injuries, smoking materials of civilian deaths and intentional fires of direct property damage.[5] Knowing the cause of fires can lead to targeted inspection, public education and code changes, and fire engineers use cause data to choose fire scenarios and estimate how often fires start.[3][6] The statistics in this article describe the United States in the years around 2000 and come from sources published up to 2008.
Elements of a fire cause
Combustion needs fuel, oxygen and heat. Van Rossum (2027) describes it with the fire triangle of oxygen, fuel and temperature. A combustible material mixed with oxygen cannot ignite unless enough heat is added, and the amount needed depends on the material.[7] The International Fire Engineering Guidelines (IFEG) note that in most environments oxygen is always present. Starting a fire then requires the accidental or deliberate combination of fuel and a source of heat, although some fires start by self-heating of the fuel itself.[8]
Hall and Cote (2008) build their account of fire prevention on three components. Every hostile fire requires an initial heat source, an initial fuel source and something to bring them together. That something nearly always has a human component, usually an immediate act or omission, or sometimes the delayed effect of an error in design or installation.[1] Prevention can act on any of the three, and causes can be defined and ranked in terms of any of them.[9][10]
Investigators analyse the ignition sequence in the same terms. Custer and Sheppard (2008) list three items. The first is the size, nature and source of the ignition energy, and the second the physical and chemical fire properties of the material ignited. The third is the set of human or mechanical circumstances that brought energy and fuel together.[2] A candidate source must be a competent ignition source, with enough energy and temperature to raise the fuel to its ignition temperature during the exposure time. An electric arc, for example, may be far hotter than the ignition temperature of a fuel but too brief to ignite it.[2]
The circumstances generally involve human factors or equipment failure. Careless welding, poor housekeeping and misuse of materials are human factors, while a ruptured valve or a failed combustion control is an equipment failure. Most equipment failures can ultimately be traced to design, manufacturing, installation or maintenance problems, or to misuse.[2] In the terminology of the US National Fire Incident Reporting System (NFIRS), "equipment involved in ignition" is the equipment that provided the heat, and it need not have malfunctioned.[11] Ahrens et al. (2008) stress that cause categories should support prevention strategies; assessing blame is a task for lawyers or philosophers.[4]
Determining the cause
In the United States, the local fire department generally has the primary responsibility to document a fire and to make an initial determination of its origin and cause.[3] If the cause is easily determined and not criminal, the investigation usually ends there. When the cause cannot be determined, a death is involved or arson is suspected, additional help is generally called in.[12] When a fire may have been set deliberately, the state or provincial fire marshal's office is responsible for the investigation in much of the United States and in all Canadian provinces.[13] The methods of Fire investigation are described in NFPA 921, Guide for Fire and Explosion Investigations.[14]
NFPA 921 was first issued in 1992 and was revised every three years up to 2004. Its 2004 edition included chapters on origin determination, fire cause determination, analysing the incident for cause and responsibility, and incendiary fires.[15] The guide also forms the basis for the performance criteria of NFPA 1033 for fire investigators, as described by Custer and Sheppard (2008).[15]
Cause determination follows a scientific approach. The investigator collects evidence, develops hypotheses about origin and cause, and tests them to see which one the evidence supports.[16] Burn patterns such as a "V", low burns or apparent multiple points of origin can be misleading, and should only be used when unambiguous or supported by other physical evidence. Fires reported to have spread "unusually fast" are often called suspicious or incendiary when the growth may have been due to the burning properties of the materials and the compartment.[2] Where a building has been totally or nearly totally destroyed, neither an on-scene investigation nor a later analysis may be able to establish origin or cause.[16]
Arson is a possible outcome of an investigation, not its starting point. Custer and Sheppard (2008) state that arson can usually be determined only after a thorough investigation has ruled out other possible causes and established physical evidence of an incendiary fire. For this reason they prefer the term "fire investigation" to "arson investigation".[17] The quality of national data depends on this work, because the determinations made by local fire officers become the data used by the national reporting system. Updates after investigation are often not entered, especially for fatal and large-loss fires.[18]
Cause categories in incident data
NFIRS records cause information in six data elements. These are a general cause category, the heat source, the equipment involved in ignition, the item first ignited, the type of material ignited, and the factors contributing to ignition.[19] Together these can define billions of distinct cause categories, so analysts use simpler groupings.[4] In Version 5.0 of NFIRS, the general cause field has the options intentional, unintentional, failure of equipment or heat source, act of nature, cause under investigation, other, and cause undetermined after investigation. Cause under investigation and cause undetermined after investigation are both treated as unknown in analysis.[20][21][4]
Intentional fires need careful reading. Older versions of NFIRS used "incendiary" when a legal decision or physical evidence showed that a fire was deliberately set, and "suspicious" when circumstances only suggested it.[11] Analysts often grouped both as arson, although NFPA 921 does not consider "suspicious" a valid cause.[22] Version 5.0 replaced both with "intentional", which does not imply an illegal act as strongly, and none of these terms corresponds exactly to arson.[11][22] An intentional fire set by a young child, for example, may not be arson in the legal sense.[11][23]
Different emphases produce different pictures of the same fires. For US home structure fires in 1999 to 2002, the general cause field put 47 percent of fires in the unintentional category. A further 25 percent were failures of equipment or heat source, and 9 percent were intentional.[23] Grouped by heat source instead, the largest categories were radiated or conducted heat from operating equipment (27 percent) and arcing (12 percent), while smoking materials accounted for 7 percent.[23] Grouping by one element is simple but often does not match the issues to be analysed. Combinations of two or three elements quickly produce thousands of categories, in which even the largest covers only a small share of fires.[4][24]
A further approach uses a small number of broad, easily recognisable categories. The most widely used scheme was a set of 13 categories developed at the US Fire Administration, which sorted each fire into one category by a hierarchy of priorities. An incendiary fire started by a cigarette, for example, counted as incendiary and not as a smoking fire.[24] Critics questioned such priorities: a fire caused by a child playing with cigarettes could be counted either as child play or as a smoking fire.[25] A revised hierarchy for the transition to Version 5.0 became less effective as more data were collected in the new version, and NFPA began to publish leading causes taken from several data elements at once. These causes are not all mutually exclusive, and the same fire can appear under more than one cause.[25][10]
Leading causes in the United States
The NFPA leading-cause tables combine NFIRS detail with the NFPA annual fire department survey into national estimates of fires reported to US municipal fire departments.[10] The table gives the leading causes of all US structure fires in 1999 to 2002, excluding the events of 11 September 2001, from the national Fire statistics. Fires are rounded to the nearest hundred, deaths and injuries to the nearest ten, and property damage to the nearest million dollars, without adjustment for inflation.[10]
| Cause | Fires | Civilian deaths | Civilian injuries | Direct property damage (million $) |
|---|---|---|---|---|
| Cooking equipment | 126,800 (25%) | 280 (9%) | 4,570 (26%) | 665 (8%) |
| Heating equipment | 65,000 (13%) | 340 (11%) | 1,470 (8%) | 867 (10%) |
| Intentional | 59,500 (12%) | 600 (19%) | 1,800 (10%) | 1,592 (18%) |
| Electrical distribution or lighting equipment | 45,700 (9%) | 240 (8%) | 1,210 (7%) | 1,087 (13%) |
| Smoking materials | 35,100 (7%) | 890 (28%) | 2,130 (12%) | 627 (7%) |
| Candle | 20,900 (4%) | 170 (5%) | 1,750 (10%) | 468 (5%) |
| Playing with heat source | 17,800 (3%) | 220 (7%) | 1,430 (8%) | 304 (4%) |
Each cause dominates a different measure of loss. Cooking equipment is the leading cause of structure fires and structure fire injuries, and it is involved in the majority of unreported home fires. Unattended cooking is the principal behavioural factor in these Cooking fires.[26] Smoking materials are the leading cause of civilian fire deaths, and most of these Smoking-material fires begin with the ignition of upholstered furniture, mattresses or bedding.[10] Intentional fires are the leading cause of structure fire property damage, accounting for nearly one dollar in every five lost.[5] Electrical distribution equipment ranks no higher than fourth on any measure except property damage, where it ranks second.[26]
Severity per fire differs widely between causes. In US home structure fires in 1999 to 2002, smoking materials caused 3.4 civilian deaths per 100 fires, intentional fires and playing with a heat source 1.6, and cooking equipment 0.3.[25] Intentional fires, candle fires and electrical distribution or lighting fires had the largest average property loss per fire, about $21,000 to $21,500.[25][27] Outside homes, intentional fires led in both the number of fires and property damage, with 18 percent of nonhome structure fires and 27 percent of their direct property damage.[28]
The ranking also changed over time. In 1980 the two leading causes of US home fire deaths were smoking and heating, and both were afterwards among the fastest declining causes. Arson and suspected arson moved past heating as the second leading cause in the late 1980s, while cooking had the least favourable trend of the four leading causes.[29] Candle fires became more prominent: their share of home fires rose from 1 percent in the early 1980s to 5 percent in 2001 and 2002.[30] Ahrens et al. (2008) note that in home fires a focus on fires points to cooking and heating, a focus on deaths to smoking, and a focus on property loss to intentional fires.[31]
Fires with unknown cause
Many fires are reported with the cause, or part of it, unknown. In nonconfined US home structure fires of 1999 to 2002, the item first ignited was undetermined or not reported in 21 percent of fires, but in 39 percent of the civilian deaths.[27] Severe fires tend to have larger shares of unknown data, because more of the evidence of origin is destroyed.[27] Since Version 5.0 of NFIRS, confined fires, such as cooking fires confined to their container, may be reported without causal data.[21] Their share of structure fires in NFIRS rose from 2.8 percent in 1999 to 34.2 percent in 2003, largely because more departments used Version 5.0.[32]
The usual remedy is to allocate the unknowns in proportion to the known cases. This assumes that the fires with unknown cause, if their cause were known, would show the same profile as those with known cause.[27] As an illustration of proportional allocation for fire counts, the allocated share of cause category is its share among fires with known cause:[27]
| Symbol | Description | Unit |
|---|---|---|
| Total number of fires in the data set | fires | |
| Number with cause unknown, undetermined or not reported | fires | |
| Number reported with cause category | fires | |
| Allocated share of cause category | – | |
| Estimated number in cause category after allocation | fires |
Ahrens et al. (2008) give a hypothetical example of two neighbouring states with the same underlying cause profile.[27] State A reports 9 percent of home fires as intentional and 14 percent with unknown cause, while State B reports 6 percent intentional and 43 percent unknown. Before allocation State B seems to have fewer intentional fires. As an illustration of the relation, State A gives and State B gives , so both states have about 10.5 percent intentional fires. The same holds for the other categories, which confirms the authors' point that both states have the same problems to the same degree.[27]
No allocation method has been shown to be best, and some analysts argue that time of day, fire size or property type could be used to allocate unknowns differently.[27] Because cause profiles differ between large and small fires, an allocation ideally takes fire size into account, although no consensus on this had formed.[27] NFPA allocates the heat source code "other open flame or smoking material" in proportion to the specific codes. Without this secondary allocation, intentional fires would overtake smoking materials as the leading cause of home fire deaths.[33] In national estimates, Hall and Ahrens (2002) describe the same allocation as built into the scaling ratios that project NFIRS data to the totals of the NFPA survey.[34]
Use in prevention and fire engineering
Cause data are used mainly to choose and design prevention. Fire cause and property type are the dimensions most often chosen in fire data analysis, because most fire prevention strategies are structured along those lines.[4] Each cause corresponds to specific prevention behaviours that education can emphasise.[35] Hall (2008) warns, however, that analysing fires only by cause can lead to an emphasis on prevention at the expense of education that saves lives once a fire has started.[35] Categories should also match the strategies available. For example, 28 percent of home fires involve cooking equipment, a more useful finding than the 34 percent that begin in the kitchen, because it locates the problem more specifically.[36]
The size of a cause category is not the only consideration. Ahrens et al. (2008) point out that a 7 percent reduction in a problem causing 500 deaths a year is worth as much as a 70 percent reduction in one causing 50 deaths. Both would save 35 deaths a year, and the authors add that some parts of the fire problem may be much more preventable than others.[36]
Fire engineers use the same elements to define fire scenarios. Hall (2002b) states that each scenario must specify the initial heat source, the initial fuel source and the circumstances that bring them together, so that fire incident databases can provide probabilities.[6] The IFEG, produced in 2005 by Canadian, US, New Zealand and Australian bodies, list ignition sources such as smoking materials, electrical equipment, heating appliances and unusual ignition sources among the hazards to review.[37] Matching measures include limiting ignition sources and separating them from fuel.[38] The IFEG also recommend treating scenarios with malicious ignition in a similar way to accidental ones. Some large-scale arson attacks may, however, be beyond the capacity of the fire safety system, so that other measures are needed.[39]
In Australia, the 2020 handbook of the Australian Building Codes Board (ABCB) on the Fire Safety Verification Method notes that combustibles in fire-isolated stairs are ignited maliciously often enough to justify a reference scenario. Its worked example treats an arson attack with cardboard boxes in a stair as a credible scenario, whose frequency security and building management could strongly affect.[40] For US nuclear power plants, the 2005 fire risk method of the Electric Power Research Institute and the US Nuclear Regulatory Commission groups fires by plant location and ignition source type. It gives generic ignition frequencies derived from a fire events database.[41] In the Netherlands, Van der Graaf et al. (2021/2023), citing earlier research, note that charging electric vehicles in indoor car parks adds new causes of fire. Examples are damaged cables or charging stations, battery management problems and collisions with charging stations.[42]
Limits of cause data
Cause statistics describe reported fires only. National estimates cover fires reported to US municipal fire departments and exclude fires reported only to federal or state agencies or to industrial fire brigades.[10] In the United States, NFIRS captured roughly one million fires a year but only an estimated half of the fires reported to fire departments. It is therefore projected to national totals with the help of the NFPA survey.[43] The firefighters who code the data are generalists: they can usually identify the equipment that provided the heat, but rarely the part involved, and brand and model are not coded.[44]
Definitions also differ between users and over time. The US Consumer Product Safety Commission excludes intentional fires from its product analyses, whereas NFPA generally includes them.[4] The NFIRS Wildland Fire Module uses cause categories that differ from those of the Fire Module, so its "equipment" cause cannot simply be combined with "failure of equipment or heat source".[45] The change to Version 5.0, which dropped the "suspicious" choice and causal coding for some fires, was expected to produce an artificial drop in arson incidents and to make trend analysis difficult during the transition.[11]
Finally, a cause category says little about blame or about the best remedy. A smoking-related death might have been prevented by more careful smoking, by bedding and furniture that resist ignition, or by cigarettes designed to self-extinguish.[4] Schwartzman (2008) notes the argument that almost no fire cause is truly "accidental". In fires started by children, for example, better supervision or storage of ignition materials can usually prevent the fire.[46]
See also
- Electrical ignition
- Heating equipment fire
- Large-loss fire
- NFPA 921
- NFPA fire experience survey
- National Fire Incident Reporting System
Notes
- ↑ 1.0 1.1 Hall and Cote (2008), pp. 3-11–3-12.
- ↑ 2.0 2.1 2.2 2.3 2.4 Custer and Sheppard (2008), p. 3-37.
- ↑ 3.0 3.1 3.2 3.3 Custer and Sheppard (2008), p. 3-31.
- ↑ 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 Ahrens et al. (2008), p. 3-71.
- ↑ 5.0 5.1 Hall and Cote (2008), pp. 3-14–3-15.
- ↑ 6.0 6.1 Hall (2002b), p. 5-147.
- ↑ Van Rossum (2027), §4.2.
- ↑ National Research Council of Canada et al. (2005), p. 3.3-2.
- ↑ Hall and Cote (2008), p. 3-12.
- ↑ 10.0 10.1 10.2 10.3 10.4 10.5 10.6 Hall and Cote (2008), p. 3-14.
- ↑ 11.0 11.1 11.2 11.3 11.4 Ahrens and Stewart (2008), p. 3-49.
- ↑ Custer and Sheppard (2008), p. 3-38.
- ↑ Custer and Sheppard (2008), p. 3-32.
- ↑ Farr and Sawyer (2008), p. 1-83.
- ↑ 15.0 15.1 Custer and Sheppard (2008), p. 3-33.
- ↑ 16.0 16.1 Custer and Sheppard (2008), p. 3-35.
- ↑ Custer and Sheppard (2008), p. 3-39.
- ↑ Ahrens and Stewart (2008), p. 3-47.
- ↑ Ahrens et al. (2008), pp. 3-70–3-71.
- ↑ Ahrens and Stewart (2008), p. 3-50.
- ↑ 21.0 21.1 Ahrens et al. (2008), p. 3-66.
- ↑ 22.0 22.1 Ahrens et al. (2008), p. 3-68.
- ↑ 23.0 23.1 23.2 Ahrens et al. (2008), p. 3-72.
- ↑ 24.0 24.1 Ahrens et al. (2008), p. 3-73.
- ↑ 25.0 25.1 25.2 25.3 Ahrens et al. (2008), p. 3-74.
- ↑ 26.0 26.1 Hall and Cote (2008), p. 3-15.
- ↑ 27.0 27.1 27.2 27.3 27.4 27.5 27.6 27.7 27.8 Ahrens et al. (2008), p. 3-76.
- ↑ Hall and Cote (2008), p. 3-16.
- ↑ Hall (2008), p. 5-108.
- ↑ Ahrens et al. (2008), p. 3-64.
- ↑ Ahrens et al. (2008), p. 3-90.
- ↑ Ahrens et al. (2008), pp. 3-76–3-77.
- ↑ Ahrens et al. (2008), p. 3-77.
- ↑ Hall and Ahrens (2002), pp. 5-67–5-68.
- ↑ 35.0 35.1 Hall (2008), p. 5-111.
- ↑ 36.0 36.1 Ahrens et al. (2008), p. 3-70.
- ↑ National Research Council of Canada et al. (2005), p. 1.2-9.
- ↑ National Research Council of Canada et al. (2005), p. 1.2-10.
- ↑ National Research Council of Canada et al. (2005), pp. 1.2-27–1.2-28.
- ↑ Australian Building Codes Board (2020b), pp. 67–68.
- ↑ Nowlen et al. (2005b), pp. 6-1–6-2.
- ↑ Van der Graaf et al. (2021/2023), p. 29.
- ↑ Hall and Ahrens (2002), p. 5-67.
- ↑ Hall and Ahrens (2002), p. 5-68.
- ↑ Ahrens et al. (2008), pp. 3-65–3-66.
- ↑ Schwartzman (2008), p. 5-87.
References
- Ahrens, M., Frazier, P., & Kelch, G. (2008). Use of fire incident data and statistics. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-61–3-92). National Fire Protection Association.
- Ahrens, M., & Stewart, S. (2008). Fire data collection and databases. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-43–3-59). National Fire Protection Association.
- Australian Building Codes Board. (2020b). Handbook: Fire Safety Verification Method (Print version 1.1).
- Custer, R. L. P., & Sheppard, D. T. (2008). Fire loss investigation. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-31–3-41). National Fire Protection Association.
- Farr, R. R., & Sawyer, S. F. (2008). Fire prevention and code enforcement. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 1-81–1-93). National Fire Protection Association.
- Hall, J. R., Jr. (2002b). Product fire risk. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 5-143–5-152). National Fire Protection Association; Society of Fire Protection Engineers.
- Hall, J. R., Jr. (2008). Using data for public education planning and decision making. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 5-103–5-118). National Fire Protection Association.
- Hall, J. R., Jr., & Ahrens, M. J. (2002). Data for engineering analysis. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 5-65–5-78). National Fire Protection Association; Society of Fire Protection Engineers.
- Hall, J. R., Jr., & Cote, A. E. (2008). An overview of the fire problem and fire protection. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-3–3-30). National Fire Protection Association.
- National Research Council of Canada, International Code Council, Department of Building and Housing, & Australian Building Codes Board. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
- Nowlen, S. P., Najafi, B., Wyant, F. J., Forester, J., Kazarians, M., Kolaczkowski, A., Joglar, F., Funk, D., Hannaman, G., & Anoba, R. (2005b). EPRI/NRC-RES fire PRA methodology for nuclear power facilities: Vol. 2. Detailed methodology (NUREG/CR-6850; EPRI 1011989). Electric Power Research Institute; U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research.
- Schwartzman, P. (2008). Juvenile firesetting. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 5-87–5-102). National Fire Protection Association.
- Van der Graaf, J., Hessels, T., & Rosmuller, N. (2023). Fire safety of indoor car parks accommodating electrically powered vehicles. Nederlands Instituut Publieke Veiligheid. (Original work published 2021)
- Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.