Fire investigation, sometimes referred to as origin and cause investigation, is the analysis of fire-related incidents. After firefighters extinguish a fire, an investigation is launched to determine the origin and cause of the fire or explosion. Investigations of such incidents require a systematic approach and knowledge of basic fire science.
Contents |
Fire investigation, like crime scene investigation, can involve several disciplines. In a typical law enforcement investigation, there will be a fire scene investigator and a laboratory analyst performing chemical analysis of fire debris. In more complex cases, there will be issues other than whether the fire was intentionally set or not. Civil cases tend to address far more issues involving fire spread, the performance of fire protection systems, and issues involving liability for the fire.
In investigations conducted by law enforcement, the major difference between fire investigation and the investigation of other crimes is that with the fire, it is first necessary to prove that any crime occurred. This is usually not the case with most criminal incidents.
There are also major differences in the qualifications of personnel investigating fire on the scene, and those analyzing samples of fire debris, or examining potential ignition sources in the laboratory. The laboratory analyst will possess at least a bachelor’s degree and a specialized understanding of the mechanics of gas chromatography-mass spectrometry.
Fire scene investigators, on the other hand, require an understanding of fire chemistry and fire dynamics, but unfortunately many field investigators do not possess such knowledge. In fact, many field investigators possess no formal education beyond high school. While there exist other forensic disciplines where technical skills learned through apprenticeship may provide adequate training (e.g., fingerprints, firearms identification, and handwriting comparison), it is difficult to argue that individuals who have a limited understanding of the chemistry and physics of fire development can draw reasonable conclusions about fires.
In recent years, increasing numbers of fire protection engineers, scientifically trained individuals with a better (but still imperfect) understanding of the behavior of fire, have demonstrated an interest in fire investigation, and the literature is beginning to reflect the influence of this group. Much of the work of fire protection engineers in this area, however, is still beyond the technical grasp of the average fire or arson investigator.
If one looks back to the mid-1970s and compares the state-of-the-art in fire investigation then with the state-of-the-art today, many important improvements will be noted. The 1980s and 90s were a time of very gradual change, as it became clear that many of the “rules of thumb” for fire investigation turned out to be based on anecdotal evidence at best and witchcraft at worst.
In 1977, the Law Enforcement Assistance Administration (LEAA) reported the results of a survey of fire investigators in a publication entitled Arson And Arson Investigation: Survey And Assessment.[1] This report contains a compilation of the “indicators” of incendiary activity used by fire investigators at the time. The authors of the survey warned that none of the indicators had been scientifically validated and recommended a series of carefully conducted experiments to learn whether these indicators were (or were not) valid. Three years later, the National Bureau of Standards (NBS-now called the National Institute of Standards and Technology, or NIST) produced a document called Fire Investigation Handbook, which repeated all of the indicators, and gave them the imprimatur of the most credible science and engineering institution on the planet.[2] Unfortunately, none of the scientific experiments required to validate these indicators had been conducted. The NBS simply took the word of two instructors from the National Fire Academy, which resulted in numerous textbooks being written with the mythology of arson investigation permanently embedded. It would take more than two decades to undo the damage.[3] In its 2009 report, Strengthening Forensic Science in the United States: A Path Forward, the Committee on Identifying the Needs of the Forensic Science Community, made the same recommendation that the LEAA had made 32 years earlier.[4] Some of that work is now taking place and will be discussed below.
In 1985, the Standards Council of the National Fire Protection Association (NFPA) became concerned with the quality of the work product of fire investigators, and commissioned the Technical Committee on Fire Investigations. The Technical Committee was asked to draft a guideline for fire investigators, and that task took seven years. The first edition of NFPA 921, Guide for Fire and Explosion Investigations, was published in 1992.[5]
To state that the fire investigation profession did not immediately embrace NFPA 921 would be a serious understatement. In fact, the howls of protest from fire investigation “professionals” were deafening. If what was printed in “that document” were actually true, it meant that hundreds or thousands of accidental fires had been wrongly determined to be incendiary fires, i.e., intentionally set. No investigator wanted to admit to the unspeakable possibility that he had caused an innocent person to be wrongly convicted, or a family to be wrongly denied their life savings. The profession was in denial.
The first serious challenge to the “old school” of fire investigators came in 1996 in Michigan Millers Mutual Insurance Company v. Janelle R. Benfield,[6] where a fire investigator who failed to properly document his observations was excluded from testifying. In the appeal of that exclusion, the International Association of Arson Investigators (IAAI) filed an amicus curiae brief, in which they contended that fire investigators should not be held to a reliability inquiry because fire investigation was “less scientific” than the kind of scientific testing discussed in the Daubert decision of 1993. Eventually though, there were enough court rulings, including the unanimous Supreme Court decision in Kumho Tire v. Carmichael,[7] to persuade the majority of fire investigators that it was necessary to accept the scientific method recommended by NFPA 921.
It is difficult to state exactly when NFPA 921 became “generally accepted by the relevant scientific community,” but 2000 was an important turning point. It was that year that the United States Department of Justice released a research report entitled Fire And Arson Scene Evidence: A Guide For Public Safety Personnel,which identified NFPA 921 as a “benchmark for the training and expertise of everyone who purports to be an expert in the origin and cause determination of fires.”[8] That same year, the IAAI for the first time endorsed the adoption of the new edition of NFPA 921.
One thing that NFPA 921 has accomplished is to make it easier to distinguish between credible investigative results and those based on hunches and feelings or discredited mythology. The Guide provides the investigator with the tools to do his job, but demands that he be able to justify his conclusions with data, sound science, and clear reasoning. This is a good result.
Some recent high-profile criminal arson cases[9] have attracted the attention of the public and the media and have resulted in some interesting studies regarding the prevalence of arson in the United States. Dave Mann of the Texas Observer became interested in the study of errors in fire investigation as a result of the cases of Ernest Ray Willis (who was exonerated after 17 years on death row) and Cameron Todd Willingham (who was executed after 12). He published a study that included a count of total fires in Texas versus the number of fires determined to be arson. Those results demonstrate a more than 60% drop in the number of fires determined to be arson between 1997 and 2007.[10] After reviewing the data from Texas, Jack Nicas, a reporter for the Boston Globe performed the same exercise in Massachusetts, with even more startling results. Between 1984 and 2008, the percentage of fires determined to be arson in Massachusetts dropped from over 20% to less than 2%, despite a net increase in the total number of fires. Nationwide, from 1999 to 2011, the National Fire Protection Association reports a drop from around 15% to around 8% in the percentage of fires determined to be arson.
At least some of the downward trend can be accounted for by fire investigators taking a more cautious approach, and being more cognizant of the consequences of their determinations. This caution is probably not the result of old-school fire investigators changing their ways. NFPA 921 has now been a fact of life for 20 years, a time period over which many poorly trained investigators have had the opportunity to retire, and new fire investigators have always been aware of the need for more caution. As the great scientist Max Planck put it, “Science advances one funeral at a time.” New ideas tend to spend a fair amount of time in the “heresy box,” and new ideas in fire investigation are no exception. When it was first posited in the mid-1980s that full room involvement could be responsible for irregular patterns on a floor, many fire investigators derided that idea as “the flashover defense.”[11] Flashover is a transition point in a structure fire from having “a fire in a room” to “a room on fire.” Post-flashover burning can create damage on a floor that looks very much like an ignitable liquid had burned there.
It was only when fire investigators themselves began allowing their weekend seminar training fires to continue for a few minutes after flashover that they began to realize what the fire protection engineers were saying was correct. The rules for interpreting fire damage change once the fire becomes fully involved. There is still a small but significant cadre of fire investigators fighting a rear guard action who refuse to accept this fact, but acceptance is coming. The best training that fire investigators receive no longer focuses on teaching them to “recognize arson,” but on teaching them how to understand fire patterns, particularly those caused by the effects of ventilation on fully involved compartment fires.
2005 marked another major turning point in our understanding of fire behavior, as well as our understanding of the accuracy of fire origin determinations. A group of certified fire investigators from the Bureau of Alcohol Tobacco and Firearms (ATF) designed an experiment that mirrored similar experiments that had been conducted (but not documented) at the Federal Law Enforcement Training Center in Glynnco, GA.
These investigators set up two rooms, simple 12 by 14 foot bedrooms, set each of them on fire, and allowed them to burn for about two minutes after they flashed over. They then asked 53 participants in a Las Vegas IAAI-sponsored fire investigation seminar to walk through the burned compartments and to write down the quadrant in which they believed the fire had originated. In the first compartment, three participants identified the correct quadrant. When the exercise was repeated on the second compartment, three different participants identified the correct quadrant.
These results caused much consternation, particularly as ATF agents began presenting the results to groups of investigators. The bottom line was “The “old-days” of locating the point of origin of a post-flashover fire by relying on the “lowest burn and deepest char” are OVER![12]
When word of the ATF experiments reached the fire investigation community, people began to examine the data more closely, because an error rate over 90% was simply unimaginable! In fact, the poor results should have surprised nobody. Carman reports that in the undocumented tests at Glynnco, the success rate was 8 to 10%.[13] Certainly, the participants in the Las Vegas tests were not allowed to interview witnesses, nor were they allowed to shovel any of the debris or perform any of the other activities besides visual observation that typically take place at a fire scene. The qualifications of some of the participants were found to be less than stellar, and some people were taking part in the experiment just to familiarize themselves with fire investigative procedures.
No matter how many explanations for the low success rate were offered, however, there was no way to increase the number of correct origin determinations beyond three. Reducing the denominator (the total number of “experienced” participants) might raise the percentage of correct answers to 10% or even 20%, but it is important to remember that 25% is the percentage of correct answers that would be expected if all of the participants were blind.
In an attempt to understand the data, Carman and his collaborators at ATF re-created the test fires at the ATF Fire Research Laboratory in Ammendale, MD, and modeled the results using computational fluid dynamics (FDS and Smokeview). What came out of these studies was a better, but certainly not complete, understanding of the effects of ventilation in post-flashover fires. The results of these studies have now been incorporated into two very well produced training modules, available at no cost at CFITrainer.net.[14]
Prior to flashover, a fire grows by involving more fuel. Once flashover occurs, all of the fuel that can be involved is involved, and the fire can only grow where it has sufficient ventilation. The fire is said to have made the transition from a “fuel-controlled” fire to a “ventilation-controlled” fire.
Fire investigators looking for the origin of a fire are typically looking for places where the fire plume has interacted with two-dimensional surfaces such as walls and ceilings. These patterns may cease to be created once the transition to flashover has occurred. In ventilation-controlled fires, patterns will be generated around doors and windows, as well as around any places where air leakage existed before the fire, or where such ventilation was created as a result of the fire. When flashover occurs, and every exposed combustible surface ignites almost simultaneously, the fire can consume almost all of the oxygen in the room. After that happens, there will be diminished burning where the oxygen is depleted. The fuel at the origin of the fire will not necessarily continue to burn, and much larger patterns not associated with the origin, can be generated. These new patterns can obscure older, plume-generated patterns, or even if the older patterns persist, the new patterns can be confusing.
When the 2005 Las Vegas fire was modeled, the oxygen concentration was shown to be very low throughout the room, except in the immediate vicinity of the doorway, and only near the floor. This oxygen concentration suggested the existence of a “floor jet.” This inflowing oxygen collided with the wall opposite the doorway and moved upward, mixing with the unburned fuel, and resulted in a heat flux of more than 150 kW/m2. Clearly, the total energy impact (heat flux multiplied by time of exposure, expressed as kJ/m2) on this surface far exceeded the total energy impact on the wall next to the origin.
The principal problem with determining the wrong origin is that the ignition source will not be found there. Finding an origin without an accidental ignition source will lead investigators who fail to understand ventilation to conclude that somebody must have placed some fuel at that origin and ignited it with an open flame. If there is an irregular burn pattern on the carpet in that area, even in the absence of a positive laboratory report, the investigator may conclude that the fire was intentionally set using a flammable liquid. Many investigators have made errors using this kind of “negative corpus” determination. Finding the correct origin the key to a correct fire cause determination, and is the most difficult part of the investigation of a fully involved compartment fire.
In 2007, ATF agents refined and repeated the Las Vegas experiment, this time in Oklahoma City. They set up three burn cells, with identical fuel and identical ventilation, but different points of origin. The cells were allowed to burn for 30 seconds beyond flashover, 70 seconds beyond flashover, and 3 minutes beyond flashover. To put these times in context, the best fire departments in big cities might have a three-minute response time. If they are not called until someone sees the fire venting out a window (a sign of flashover) the chances of them extinguishing the fire with less than three minutes of post-flashover burning are practically zero. The results of the Oklahoma City experiment validated the data from Las Vegas obtained two years earlier. Further, it became clear that the longer the fire was allowed to burn after flashover, the less likely the fire investigators were to correctly identify the quadrant of origin. The results of the Oklahoma City experiment are shown in Table 1.
| Time | # responses | # correct | % correct |
|---|---|---|---|
| 30 sec | 70 | 59 | 84 |
| 70 sec | 64 | 44 | 69 |
| 3 min | 53 | 13 | 25 |
Table 1. Results of 3 burn cell tests conducted to measure fire investigators’ ability to determine the correct quadrant of origin[15]
There were, apparently, six investigators who ruled the origin “undetermined” based on the fact that they did not turn in a response for the 70-second post-flashover fire, and 17 investigators who declined to select a quadrant of origin when the fire had burned for 3 minutes beyond flashover.
Of those 53 investigators who did respond, the twenty-five percent (25%) who got the quadrant of origin correct. While this is a better result than the 6% obtained in Las Vegas, it was no better than would be expected if the investigators had chosen the quadrant of origin at random. Further, there are those who would argue that 69% correct or even 84% correct are low numbers, when one is using those determinations to either send people to prison, or to deny them coverage under their insurance policy.
What these results show is a failure of the infrastructure for training fire investigators. Exercises conducted at fire investigation seminars historically have used short-lived fires, extinguished before flashover, to help investigators “recognize arson.” This kind of training is no longer acceptable. What these results also show is that fire investigators and the people who employ them need to be prepared to accept the reality that sometimes the best answer that can be obtained is “undetermined,” if either an accidental or an incendiary call is not supported by conclusive evidence.[16]
In the United States, fire investigators often refer to NFPA 921: Guide for Fire and Explosion Investigations (National Fire Protection Association).
Also, Kirk's Fire Investigation by John D. DeHaan and David J. Icove has long been regarded as the primary textbook in the field of fire investigation.[17] It is currently in its seventh edition (published in 2011, ISBN 0-13-508263-3).
Fire investigators conduct their investigations using a systematic approach utilizing the scientific method, including the following[18]:
The investigator:
Spoliation
Spoliation is the destruction or alteration of evidence through intention or ignorance. The mere act of extinguishing a fire can destroy potential evidence of arson or what is also known as an "Incendiary fire." Firefighters are educated that the stream of their fire hose or the use of a Pike Pole can destroy evidence and efforts are made to do what is required to extinguish the fire, while not destroying clues to the fires' origin. A fire investigation was once compromised by a fire fighter turning off the knobs on a gas stove in the interests of safety after a house fire was knocked down. In the following investigation the homeowners daughter was accused by her father of leaving the stove on after she left the house but there was then no way to accurately determine the position of the burner knobs on the stove. Though there were no criminal issues involved in this fire, this incident of spoliation created a lack of closure for the family and feelings of distrust and animosity within the family members. -FIRE INVESTIGATION by Russell Chandler. Delmar Cengage Learning - Publishers.
The National Fire Protection Association (NFPA), through a document known as NFPA 1033, Standard for Professional Requirements for Fire Investigator, publishes minimum requirements for the knowledge skills and ability of a fire investigator. Principal among these is a 13-point list of areas in which a fire investigator is required to have education beyond high school level. These 13 topics are:
1) Fire science (2) Fire chemistry (3) Thermodynamics (4) Thermometry (5) Fire dynamics (6) Explosion dynamics (7) Computer fire modeling (8) Fire investigation (9) Fire analysis (10) Fire investigation methodology (11) Fire investigation technology (12) Hazardous materials (13) Failure analysis and analytical tools
Fire scene investigators may become certified through the International Association of Arson Investigators (IAAI) or the National Association of Fire Investigators (NAFI). Both certification programs rely heavily on the content of NFPA 1033 and NFPA 921. Both also require an application process detailing the investigator's education, training, and experience, and successfully challenging a written examination. Certificates are valid for a period of 5 years, at which time an investigator must demonstrate continued participation in the field and a minimum amount of continuing education in order to be recertified.
The International Association of Arson Investigators (IAAI), a professional group of fire investigators, grants the following certifications: Certified Fire Investigator (IAAI-CFI) – certified by the ProBoard Fire Service Professional Qualifications System. Fire Investigation Technician (IAAI-FIT) Certified Instructor (IAAI-CI) Evidence Collection Technician (IAAI-ECT)
The National Association of Fire Investigators (NAFI), a professional association of fire and explosion investigators, offer several National Board Certified fire investigation certifications including: