Editor’s Note: The paper on which this article is based was originally presented at the 2019 IEEE International Symposium on Product Safety Engineering in San Jose, CA. It is reprinted here with the gracious permission of the IEEE. Copyright 2019 IEEE.
Every forensic investigation requires an objective; in every fire investigation, that objective is to determine the origin and cause of the fire; said another way, to explain how a competent heat source came in contact with the first fuel ignited. Occasionally, “domain irrelevant” information is encountered regarding circumstances not specifically relevant to the origin and cause investigation [1].
The fire occurred, but nobody knew why. The homeowners said they thought it was a Christmas tree fire; the public agencies accepted that as plausible and likely and saw no reason to investigate further. Routine insurance company checks revealed a series of “red flags” (information that was domain irrelevant to the science-based part of the investigation); they opted to investigate more deeply and retained a private fire investigator. The explanation accepted by the public agencies, that the fire was an accidental Christmas tree fire, was accepted as a testable hypothesis by the private investigation team. The next step was to search the fire scene for data (physical artifacts) to test this hypothesis.
In addition to the original fire investigator, the insurance company retained the services of an electrical forensic engineer, a chemist, and a wood specialist to provide the science-based approach to the collection and evaluation of the artifacts and information discovered throughout the course of the investigation. This science-based approach is called the “scientific method” and has been clarified and expanded on in every edition of National Fire Protection Association (NFPA) 921 since 1992 [2]. If followed properly, this method will prevent the effects of bias and help to ensure that conclusions drawn are based on legitimate data and have withstood rigorous scrutiny and testing.
Investigative Process
The scientific method is based on the sequential process of gathering data, analyzing the data, and forming one or more hypotheses. The hypotheses are then tested in an attempt to disprove the theories. Many times this requires the collection of additional data, the application of the laws of physics, or a comparison to other accepted hypotheses. The selection of the final hypotheses may require multiple iterations of the cycle.
If the scientific method is followed, the fire investigation will proceed in a methodical fashion. For fire investigation, the first step is to determine the origin of the fire. Only after the origin has been accurately determined can the cause of the fire be hypothesized. Note that if the origin has not been correctly identified, there is very little possibility the correct cause will be determined. The cause must consider such things as competent ignition sources, potential first ignited fuels, and the combination of the two resulting in a combustion process that can continue.
Only after the correct cause has been determined (using the scientific method) can responsibility for the fire be considered. Information regarding responsibility is not relevant to the origin and cause determinations and the premature disclosure of such information can introduce bias in the remainder of the investigation. If used correctly, the scientific method will compensate for any irrelevant knowledge provided to an investigator.
Structure Description
The structure was a 1940s vintage mountain cottage (Figure 1). The original exterior construction was of vertically set railroad ties insulated with newspaper. The building had at least one add-on, apparently to add a separate kitchen room and a deck. The interior was finished with modern gypsum board and a metal roof had been added. The house had been lifted adding a crawl space which provided a means for an electrical service upgrade to use modern circuit breakers.
Origin Determination
Potential Ignition Sources
The scene was examined on several occasions between December 2007 and May 2008 (Figures 2 and 3). Every witness statement was treated as a hypothesis, and physical artifacts were collected for laboratory examination. One difficulty in this investigation was the rich mix of domain relevant and irrelevant information provided to the fire investigation team before the cause investigation was completed. Special care was required to prevent any bias from slipping into the investigation.
The area of origin for the fire was determined by the fire investigator based on his interpretation of the fire patterns. Through his interviews with the homeowners, the investigator learned of the presence of road flares and their approximate location, as well as the location of the Christmas tree; he was able to recover the road flare ash residue for chemical analysis. He was not able to immediately rule out electrical devices, such as electric Christmas lights, timers, or surge suppressing outlet strips, as potential ignition sources.
Fire Patterns
A fire pattern is any measurable effect that results from a fire. Fire destroys evidence, but it leaves its own kind of residue, and persons so trained can identify the origin of the fire. An experienced investigator can recognize what the damage patterns reveal in terms of the sequence, development, and progression of a fire [3].
Fire Dynamics
A heat source must be hot enough, continue to be hot long enough, and be close enough to transfer its heat to the first fuel ignited, and that heat must be sufficient to raise that fuel to its autoignition temperature. For the first fuel to continue to burn, The heat released must heat an adjacent fuel to its ignition temperature. Lastly, there must be enough oxygen (air) to combine with the burning fuel. Fuels, ventilation, fire dynamics, and suppression are but a few of the considerations in the origin determination.
Scene Processing
The final scene processing and collection of physical artifacts took place in May 2008 (Figure 3). Most of the interior walls and framing were burned down to the wall sill plates. A three-foot-by-four-foot section of floor was cut and removed in its entirety, including the Christmas tree stand, electrical power cords, and tree lights. Ash had sifted through holes in the floor in this area and had formed conical piles about six to eight inches in height. The location of the ash piles was documented on the scene diagram and each ash cone was sampled in three layers. The cut floor opening allowed free access to the ash piles from above and enabled the ash piles to be collected using proper technique to avoid cross-contamination of any potential hydrocarbons.
Fuel and Heat Source Analysis
Twenty distinct potential heat sources and fourteen potential first fuels were reviewed; this amounted to considering 280 separate hypotheses, each of which was tested and ruled as possible or eliminated. The homeowner’s hypothesis of ignition and first fuel (decorative lights and cut decorative branch swags) was ruled out by testing exemplars. The lights did not reach sufficiently high temperatures to ignite lightweight cellulose material such as tissue paper. Other hypotheses were similarly tested using structural fire patterns, electrical fire patterns, the location and condition of potential heat sources, and identification and location of the potential first fuels.
Five potential first fuels were close enough to the area of origin that no single fuel could be identified as the first fuel; however, one was identified as most likely. Chemical analysis of the road flare residue, and several samples from the sub-floor ash cones, showed substantial amounts of medium petroleum distillate (MPD). No source was found that could explain either the extreme amount of MPD or its presence in the precise location both above and below the floor and in the flare ash.
Data Analysis and Hypothesis Testing
Timeline
Typically, the occupants of a home are interviewed early on in a fire investigation. Useful information (not data) may be offered regarding nuances with the residence that may assist the investigator in their origin and cause determinations. Two kinds of events appear in a timeline: hard events (those that can be corroborated with hard copy, generally machine recorded timestamps), and soft events (such as estimates and witness recollections). A fire timeline starts with the fire report that shows the call time, whether it was a 911 call, the dispatch time, or the arrival time of different suppression personnel.
Homeowner statements provided information that could be placed in a chronological sequence. Such timeline information can be verified against other hard data. The timeline information leading up to and at the time of ignition are of great value in determining the competency of an ignition scenario. The laws of physics are invoked to establish cause and effect, what is possible and what is not. Some of the timeline information was relevant to the origin and cause with respect to early fire development.
Electrical Heat Sources
During the scene processing in May 2008, the locations of the outlets had to be determined from pre-fire photographs and the locations where the branch circuits penetrated the sill plates from below. The most recent wiring branch circuits had been routed below the floor and up into the walls. The wiring could be accessed from the crawl space. Inspection showed several generations of wiring, indicating that the cottage had been rewired several times over the years. Only one light switch and outlet were in the area around the Christmas tree. No timers were reportedly in use, and none were found. The decorative lights were plugged into a non-surge suppressing outlet strip which had a metal enclosure.
Examination of the homeowner’s exemplar lights showed that they were in good condition. The incident tree lights were badly fire damaged; however, no differential damage or electrical arcing was found to indicate a failure.
Road Flare Residue
The fire investigator found residue of the homeowner’s reported road flare approximately where it was observed in the homeowner provided photographs. The residue was sent for chemical analysis which confirmed that it consisted of road flare and “substantial amounts of MPD.” Review of the chromatograms also showed the presence of traces of gasoline.
Chemical Analysis
On the May 2008 visit to the fire scene, five months after the loss, additional samples were collected from ash debris in the crawl space below the area where the tree was located. These samples were sealed in vapor-tight containers and went untested until two years after they were collected. These samples also showed extremely high levels of MPD. We now had more lamp oil fuel, and in a location where we could not explain its presence.
Witness statements could not be completely disregarding during origin determination; additional inquiries were made because the MPD could have had reasonable accidental explanations. Early analysis of the road flare ash revealed high concentration of MPD. Since this was unexpected, the homeowners were asked if they had fuels such as lamp oil in the area. They said they had some in several emergency hurricane lamps that “must have collapsed and fallen in the area of the tree.” Metal components from the hurricane lamps were in fact found, but at a substantial distance from where the MPD was found. Pre-fire photographs confirmed the reported pre-fire location of the lamps, but also showed that the volume of oil was likely insufficient to account for the high quantities found post-fire.
Upon further inquiry, the homeowners indicated that several quart or half-gallon containers were stored under the stairs, adjacent to the tree, and also in or near the entertainment center, on the opposite side of the tree from the stairs. Such containers burn at the top, but the bottoms normally remain intact on the floor, often protected by fire debris that covers them. No such containers were found. Samples from these locations were collected and tested. No containers and no MPD products were found in these locations.
Photographic Metadata
Additional data was recovered from the metadata in the images provided by the homeowners. Modern digital camera equipment creates a file encoded with bits that represent points of color and light and dark contrasts in an image. The cameras then attach non-image information to the file that includes the date and time of the photograph, the camera settings, and the camera make and model information, among other things.
The homeowners provided a CDROM of photographs taken before, during, and after the fire. The provided photographs were ten-megapixel images; however, in 2007 there were no ten-megapixel cell phones thus the photographs were not taken with the homeowner’s cell phone as was claimed. Scientifically determined and factual but domain irrelevant to origin and cause; the information was forwarded to the insurance company.
Photographic Information
Included on the CDROM were outside photos of the house (north and south), inside photos of the tree and decorations from multiple angles (showing the tree, cut swags, lights, candles, and candelabras on the upright piano), gifts under the tree, and yes, the wrapped road flares are visible literally front and center.
The first series of photographs showed the exterior of the house and yard; the images are time stamped December 19 and December 22 (the day of the fire). Photos taken prior to the fire confirmed the locations of the tree, the decorations, and the gifts under the tree, including the road flares.
The homeowners also supplied photographs that included images from around the time of the fire. The images span about five minutes showing the house fully engulfed in flames with fire plumes from every window and door (Figure 4). Then there is a gap of about eight minutes—the 911 call time is in this window of time. The photographs resume a few minutes after the 911 call; they show more photographs of the house engulfed in flames. The homeowner’s statement that he went to the south end of the house to rescue his wife was a testable hypothesis. The absence of footprints in the snow on that end of the house demonstrated that his statement was false.
FDS Simulation
The pre-fire photographs supplied by the homeowners provided information regarding the dimensions and structure of the house. Fire Dynamics Simulator (FDS) version 4 is a computational fire-driven fluid flow modeling software package publicly available from the National Institute of Standards and Technology (NIST) [4]. FDS allows a knowledgeable person to construct a virtual representation of a building replete with combustible and non-combustible materials. FDS models the process of combustion and will generate a very good approximation of combustion products, oxygen consumption, combustion temperatures, flame plume and hot gas buoyancy, and soot production.
Simulation was performed to verify and create a rendered visualization of the ignition, combustion, plume development, approximate temperatures, and fire development timeline of a burning tree. The model was reviewed and compared to the fire damage patterns to ensure that the fire dynamics, chemistry, and physics of the fire were consistent with the model. The simulation also provided validation in determining the extent that a full-scale mock-up would be required. The model showed that fire plume development would propagate, as expected, up the staircase (Figure 5).
Botanical Analysis
One of the problems with test burns (described below) is that they are normally challenged with regard to the degree of exactness with which they represent the original fire conditions. While not all variables can be known or controlled, the most significant variables are identified, and those materials and conditions are selected and controlled for the test.
Botanical identification was used to verify the homeowner’s recollection of the tree age, size, and species. The tree was substantially consumed in the fire; however, the base of the stump was recovered in its stand and the growth rings of the tree were recognizable. Needles from the fire debris were recovered and cross-sectioned. The cross-sectional appearance of a needle is unique to the species. The incident tree was identified as a noble fir, confirming the homeowner’s recollection. Five noble firs of the same age as the incident tree were purchased.
The wood specialist confirmed that the species of tree and the degree of watering (or desiccation) have a large effect on the combustibility of a particular tree.
Burn Cell Test
In the world of fire investigation, it is common to construct rooms with dimensions representative of actual rooms, and containing fuels for the purpose of validating one or more hypotheses. Such rooms are referred to as “burn cells.”
The cell is instrumented with cameras and thermocouples. An ignition source is then used to start a fire. The fire is extinguished at some pre-determined time or event, and the resulting damage is analyzed. The analysis consists of using the resulting fire patterns to compare with the incident fire damage patterns in order to test a fire ignition and development hypothesis.
The date of the test burn was selected, and the trees were decorated with lights similar to those used by the homeowners. The trees were watered for one, two, or three weeks, then not watered until the day of the test burn.
A fire restoration company was retained to construct a mockup of the staircase, archway, and a portion of the living room using conventional construction materials (Figure 6). The interior walls were finished with five layers of 1/4-inch gypsum wall board so that after each test burn the damage patterns could be photographed and the damaged layer easily removed to prepare for the next test burn. One day was required to construct the burn cell and five burn tests were completed on the second day.
Test burns were conducted at a local fire department training facility. Fire department personnel were invited to witness the tests, and several were on hand in turn-out gear, in the event that the tests produced more vigorous fire development than expected.
Christmas Tree Fires
It is a common misconception that Christmas trees are responsible for many holiday fires, this is simply not born out in the published literature. From 2011 to 2015 the NFPA reported an average of 200 Christmas tree fires per year [5]. Of those, 36 were reported as caused by lights, 26 were reported as caused by wiring or related equipment, 12 were reported as caused by cords or plugs, and 48 were reported as arson. NFPA did not specify whether the fires occurred in natural or artificial trees. Over the same period of time, an average of 27 million real Christmas trees were sold per year [6]. Christmas tree fires are rare these days—perhaps fires were more common when flaming candles were used to decorate the trees.
Tree Ignition Hypotheses
The tests were performed with the most well-watered tree first, since it was expected to produce the least damage to the test cell. Had the most well-watered tree accelerated to full combustion, the homeowner’s statements would have taken a different weight and the case may not have gone to trial. Had the fire accelerated quickly, the homeowners may not have escaped the house at all, and the homeowner’s claim that he was at the top of the stairs witnessing a flickering on the wall would make no sense since the hot combustion gases rising to the top of the stairs would have overtaken him.
If two or three of the most desiccated trees burned well and the remaining most watered trees did not, again the insurance company would likely not have taken the case to trial because a tree-fueled fire would have been shown to be a physically possible scenario.
If the tests showed that all of the trees were poor fire accelerants, then some other mechanism would be required to cause the fire to develop more forcefully. The accelerating effect of lamp oil on carpet, ignited by a road flare, could produce a heat release rate sufficient to take the gifts and the tree to full combustion.
Test Burn Results
Noble firs are extremely difficult to ignite, they do not burn like Hollywood special effects nor do they burn like the test subject trees found on the NIST website. It turns out that we were unable to ignite any of the four most watered test trees. When ignition did occur, the fire self-extinguished. A burning matchbook was used as the ignition source (to represent an electrical hot spot); it ignited adjacent needles and stems but they did not produce enough heat to ignite the adjacent high moisture content needles and twigs. Only about one-third of the most desiccated tree burned before self-extinguishing. The tests showed that this species of tree was a poor first fuel.
This result indicated that a more substantial fuel source of longer duration, was required to heat a noble fir to bring it to fully involved combustion, even after five weeks of desiccation. A final test was devised and tested. A square section of carpet was placed under the last (most desiccated) tree. About 500 ml of lamp oil was added to the carpet, prior to ignition with a torch. The wicking effect of the carpet produced a generous plume that needed to be extinguished by the fire department.
Homeowner Statements
Insurance Special Investigation Unit (SIU) investigators had information that was not shared with the fire investigation team, and the information acquired during witness interviews was carefully separated to provide only those statements that were testable and relevant to determining the origin and cause of the fire.
The fire investigator in this case was an experienced interviewer and had received training over the years to interpret interview statements. Questions during an interview normally start with a broad perspective, such as How old is the structure?, How long have you lived here?, When was the shop building added?, Did you make any renovations, remodels, or wiring changes?, Were there any electrical problems, flickering lights, switches, outlets, or appliances that snap, crackle, or pop?
Such questions are important because the homeowners know the contents of their residence and the age and condition of the appliances. The homeowner’s opinions are typically domain irrelevant to origin and cause but must be considered as testable hypotheses. Such statements normally identify potential fuels and potential heat sources that can be itemized then identified during the physical processing of the fire scene. The interviewer must remember that the responses are simply hypotheses that must be tested prior to acceptance.
An example of one follow-up topic from this investigation went as follows:
What do you think caused the fire?
“We think it started in the Christmas tree.”
Out of curiosity what makes you say that?
“Well, we looked in the window and the Christmas tree was on fire.”
How was the tree decorated, when did you buy the tree, what kind of tree was it, how was it watered, did it seem dry, were the needles falling off, where was it located, did you cut it or harvest it, where did you buy it, what kind of lights did you have on it, where were the lights plugged in, were the lights on a timer, what condition were the lights in, were they on at the time of the fire?, etc.
These are all physical artifacts that can be identified as the fire scene is processed.
Some of the responses struck the investigators as odd. Were there gifts under the tree? “Yes.” What kind of gifts? “Wrapped gifts.” Were they in boxes? “Yes, wrapped boxes.” What was in the boxes? “I think there were clothes, (he looks at her, she looks at him) yes, there were clothes.”
The witnesses are not as forthcoming to these questions, and their answers were lacking in the former amount of detail provided to other questions. Were they looking at each other because they didn’t know what the other intended as a gift, or were they wondering if they were giving something away by answering the question? Again, these are domain relevant to the science-based investigation only to the extent that they name physical artifacts that can be identified during scene processing. Other details may be of value to the insurance SIU investigation; it is important to let the fire investigator take the lead with interviews as he/she is the one who has had training in reading between the lines.
Continuing the questions: Was there anything else? “Yes, I think there was some electronics, (he looks at her, she looks at him) yes, there were electronics.” Was there anything else? “Yes, there were some road flares.” Tell us about the road flares. “They were wrapped in groups of three, we bought them at Bi- Mart, on sale. We wanted inexpensive, practical gifts for our church friends.”
Do you have any photographs from before the fire? “Yes, I love taking photographs especially this time of year, everything is so beautiful.” What kind of camera do you have? “I used my cell phone.”
Once the structure is covered, the next level of detail has to do with things relevant to the fire. The interview moved into the next phase: What did you do the day before the fire? What did you do the night of the fire? Who discovered the fire? Describe what you did and saw, just before discovering the fire. Were you home at the time? What activities were you engaged in? What did you see, hear, or smell?
Answers are naturally used to develop additional follow-up questions in order to obtain more specific information. A tremendous amount of detail was provided to that last question and provided many hypotheses to be tested:
“My wife went up to bed, I told her I would be right up. I went up and she was already asleep. A few minutes after going to bed, I heard a weak chirping from the smoke detector, not a full alarm. I went to the top of the stairs, I smelled something like pine incense and saw a flickering on the wall and I realized the tree was on fire.
“I woke my wife and told her to get out. I went out through the window in my radio shack, I’m a ham radio operator. I thought she was behind me, so I climbed down my radio tower, realized she wasn’t behind me, I climbed back up and tried to get back in through the window, but it was too hot; that’s how I burned my hand [(Figure 7)].
“I climbed back down and ran around to the other end of the house calling for her. She had her head stuck out the bedroom window shouting for me to help her, that she was on fire. She climbed out on the roof and I told her to jump and I would catch her [(Figure 8)]. She jumped and we both fell down in the snow. I put her in the car to stay warm and I walked to the neighbor’s house to call 911.”
Some hypotheses can be tested by thought experiments. One such test of the above statement includes the fact that the wife weighed an estimated 140–150 pounds. The husband’s description of her jumping from a window 15 feet in the air into his arms, complete with his self-described bad back, with no lingering injuries, seems unlikely after looking at the photograph. Climbing out of the window, sliding down a metal roof, and climbing down the ladder isn’t as dramatic but seems more plausible—but this was not their story. Witness statements may be factual and accurate and then again, they may not. What they undoubtedly provide is an excellent field of rocks to look beneath, a rich collection of events that can be tested with physical evidence.
From this conversation, many statements presented scenarios that needed to be tested using physical evidence. Actual burn injuries, photographs of the yard around the potential exits from the residence, first responder observations, etc.; all were provided to SIU investigators, but were not generally relevant to the origin investigation.
Non-Cause-And-Origin SIU Considerations
What information did we gather that could be tested? We can examine the details of the burn injuries; the photographs that do not show the footprints in the snow; there is enough information to perform a fire dynamics simulation pursuant to a full-scale test-burn mockup; we know the tree age and species, the kind of lights on the tree (the homeowner had exemplars), and the type of road flares under the tree (another potential ignition source and/or fuel); we know that the tree was purchased pre-cut; and we know where the tree stand was located in the house. We also know that they have hurricane lamps on top of the entertainment center and that they are filled with lamp oil. Lastly, we know that first responders found the homeowner outside fully clothed while his wife was in a nightgown.
Public Agency Information
Fire Department EMT and hospital staff reported slightly different observations than was told by the homeowners, perhaps different parts of the same tale; regardless, the takeaway is that the husband had burns on his right hand, and the wife had burns on her back. First responders said that the homeowners reported that they had a file box full of business records blocking the upstairs door, that they ran down the stairs and out the front door. Realizing they had forgotten the file box, the wife ran back in to get the box; however, the fire was well developed enough that she had to exit the house. We do not know which door they exited.
The husband was right-handed and his explanation for his burn was that it happened when he was climbing the tower to go back into the window. However, looking at the photograph this story seems like a stretch. The photos showed the location of the radio tower outside the gable window of the radio shack; it was to the right of the window as one faces that end of the house (Figure 7). The natural way to re-enter was to climb up the tower, step onto the roof, and reach for the window frame with the left hand while holding the tower mast with the right hand. The EMT report of them saying they left by the front door seems more plausible—however, the homeowners would have had to come up with a different story to explain their burns.
Insurance “Red Flags”
Every viewer of television crime mysteries knows about means, motive, and opportunity—this is not just a movie trope. Whether a loss is criminal or civil, the arsonist has 1) a reason to set the fire, even if it may be as frivolous as thrill-seeking. The suspect must also have 2) the means—that is, access, knowledge, and/or the materials—to set the fire. And the suspect must also have 3) opportunity—that is, suitable circumstances to permit them to bring the ignition source into contact with the first fuel ignited. Insurance company SIUs, as well as public agency police detectives, both look for these three criteria.
Insurance companies also have access to other information that may cause them to question a claim. A sudden dramatic increase in the value of an insurance policy shortly before a fire, and/or financial distress, as in a business down-turn, property tax burden, or delinquent mortgage payments, may provide a motive for those financially desperate. It turns out that the homeowners had all these red flags. They are not proof, but they strongly suggest looking at other factors including the extensive investigation conducted in this case.
Financial Motive
Since the physical location of the origin of the fire had been identified, and accidental causes had been eliminated, the investigation progressed to the question of responsibility. The insurance company’s investigation was wholly separate from the forensic investigation. A change in the policy coverage shortly before a fire loss, added together with evidence of financial distress create a scenario where “arson for profit” becomes a reasonable motive. In this instance, the insurance company had additional information that indicated that the homeowners had inflated their claim by overestimating the value of their possessions, as well as moving low-value items into the premises prior to the fire. For example, the homeowners were not piano players, but they purchased a non-functional upright piano for $25 and claimed the value at $7,000.
All of these indicators were present; however, not all of this was known by the origin and cause investigators, and what information may have been known was irrelevant to the science-based origin and cause investigation.
Trial
The normal expectation would be that the insurance company would deny the claim and then wait for repercussions; however, in this case the insurance company preemptively sued the homeowners. Obviously, the insurance carrier did not make the decision to incur the expense of a trial either quickly or casually: trials are expensive and both parties are generally risk-averse to the potential of losing.
The fire happened in December 2007, and the detailed investigation occurred in May 2008; however, the case did not go to trial until April 2012. The trial was one week in duration. Thirteen jurors attended, twelve and an alternate.
The alternate was dismissed on Thursday but returned on Friday to hear the closing arguments and final verdict. We spoke with her after the trial, “We were ready to give them the benefit of the doubt, but when we started seeing the evidence on Wednesday, we knew they did it.” With a huge potential for small town sympathy and local camaraderie, the jury verdict was unanimous in favor of the insurance company.
Takeaways
- Good investigation requires that you actually look; if we knew what happened, we wouldn’t call it investigation.
- Use of the word “forensic” is not just for the criminal investigation world, it is for any matter involving detailed technical investigation that has potential civil or criminal legal ramifications.
- What events can happen, or not happen, rests on physical principles that fire investigators and other forensic practitioners are specifically trained to apply and interpret.
- Deliberate effort must be made to avoid being biased by domain irrelevant information encountered during an investigation.
- Fire debris artifacts can be, and should be identified, preserved, and analyzed, possibly well after the fire loss.
Myths
- It is common to think that a fire that burns a two-story house to the ground is impossible to investigate; they are difficult but not impossible. Had it not been for the insurance red flags, investigation may not have gone forward. Do not give up too soon.
- It is common to think that Christmas trees are a highly competent and probable first fuel; the published literature and the test burns for noble fir trees dispel those as myths.
- It is common and not unreasonable to think that if accelerants such as road flares or lamp oil were used, that such a completely burned structure, with a long delay in suppression would have destroyed all evidence of those combustible materials. The fire may destroy those things, but not as often as some think—it requires investigation. We don’t know what we’ll find until we look with skilled eyes.
- Again, it is common, and not unreasonable, to think that volatile liquids will not persist after five months of exposure to unlimited weathering. Yet, saturated amounts of medium petroleum distillates were, in fact, recovered, preserved, and detected in unambiguous high concentrations.
Acknowledgment
Our thanks to the insurance carrier and legal counsel who allowed us to report this cold case, albeit thoroughly redacted, and who must also remain unnamed. We also wish to thank Joanne Noel (ESi) for the idea for the title of the paper, and to Sara Kenley (ESi) for her scrupulous proofreading and editing.
References
- Forensic Science Standards Board, “OSAC Research Needs Assessment Form,” 2015.
- NFPA 921, Guide for Fire & Explosion Investigations, 2017 ed.
- NFPA 1033, Standard for Professional Qualifications for Fire Investigator, 2014 ed.
- National Institute of Standards and Technology (NIST) and VTT Technical Research Centre of Finland, “Fire Dynamics Simulator (FDS) and Smokeview (SMV),” accessed February 10, 2019.
- NFPA, “Home Christmas Tree Fires: Fact Sheet,” 2017, https://www. nfpa.org/-/media/Files/News-and-Research/Fire-statistics-and-reports/Fact-sheets/ChristmasTreeFactSheet.pdf.
- Statista: The Statistcs Portal, “Christmas trees sold in the United States from 2004 to 2017 (in millions),” accessed February 11, 2019.
Louis Bilancia is an electrical engineer with 14 years of experience in biomedical design of implanted pacemakers and defibrillators and with 11 years in the design and programming of industrial laser control, lumber sorting, and precision electronic instrumentation. He specializes in fire origin and cause evaluations, including those where electrical activity may have been involved, and has performed nearly 600 fire and electrical investigations. Bilancia can be reached at lfbilancia@engsys.com.
Dale Mann is a certified Forensic Chemist with over 30 years of experience performing a wide variety of forensic analyses in support of criminal and civil investigations. He has worked on fire and bombing investigations since 1981, and is certified both as a fire origin and cause investigator and laboratory chemist. He regularly assists criminal defense councils in reviewing prosecutorial investigations and trials. His areas of expertise include scene documentation, laboratory analyses and scenario testing. Mann can be reached at dcmann@engsys.com.
