Have you tested your theory? Your risk of exclusion just doubled.

Gerald A. King, J.D., CFEI
Armstrong Teasdale LLP, USA

Karrie J. Clinkinbeard, J.D., CFEI
Armstrong Teasdale LLP, USA

Presented at International Symposium on Fire Investigation, 2012


Both federal and state courts require experts to employ reliable scientific methodologies. Various jurisdictions employ different factors in determining whether an expert’s methodology is scientifically reliable. For years, courts have excluded unreliable expert testimony in fire litigation, an issue previously the subject of Experts Beware: Ignoring the Scientific Method Can Be Hazardous to Your Testimony , an article and presentation included in the ISFI 2010 Proceedings.

Despite the fact that the courts and NFPA 1033, the Standard for Professional Qualifications for Fire Investigators, require fire investigators to employ all elements of the scientific method as the operating analytical process throughout the investigation and for the drawing of conclusions, experts still fail to utilize the scientific method to ensure the reliability of their opinions. Some experts still believe that they can ignore the scientific method by claiming NFPA 921 is only a guide. Although many courts cite NFPA 921 as the “industry standard” for judging an expert’s methodology, other courts exclude experts for the failure to follow the principles articulated in NFPA 921 without ever citing to NFPA 921. An expert’s claim that NFPA 921 is “only a guide” does not relieve the expert from demonstrating that the chosen methodology is scientifically reliable. The legal standards governing the admissibility of an expert’s opinion demand a scientifically reliable methodology. Ignoring the principles set forth in NFPA 921 can and with almost certainty will result in exclusion  of an expert’s opinions (in whole or in part) and open an expert to the risk of third party lawsuits.

This article provides an updated look at these issues and analyzes recent cases where an expert’s testimony has been excluded in fire litigation and, in some instances, the expert has been sued for these deficiencies. Experts are excluded for ignoring facts or relying upon speculative facts. There is a trend towards excluding experts who do not test their ignition theory, particularly in product liability cases. The common theme in these cases continues to be that the expert ignored a step in the scientific method and, therefore, the opinions are unreliable and inadmissible.

Download the complete paper

Logging accident sparks forest fire in Oregon Coast Range – August 14, 1933

1933, Aug 14:  Logging accident sparks forest fire in Oregon Coast Range

On August 14, 1933, a devastating forest fire is sparked in the Coast Range Mountains, located in northern Oregon, 50 miles west of Portland. Raging for 11 days over some 267,000 acres, the blaze began a series of fires that struck the region at six-year intervals until 1951 that became known collectively as the Tillamook Burn.

The first Tillamook Burn fire—which began around noon on August 14, 1933—was sparked in a logging operation located on the slopes above the North Fork of Gales Creek, west of the town of Forest Grove. An official investigation of the fire found that it stemmed from friction produced when loggers dragged a large Douglas-fir log across a downed tree, igniting a large amount of logging debris in the area. Weather conditions—including an unusually high temperature of 104 degrees Fahrenheit, with only 20 percent humidity—helped ignite and spread the blaze, and within an hour, the fire had destroyed 60 acres of the surrounding land.

Present-day roads and highways in the region had not yet been built, and the remote location of the logging operation meant that the loggers were forced to fight the fire largely by themselves. Some 3,000 men, including loggers, local farmers and volunteers and several hundred members of the Depression-era Civilian Conservation Corps, battled with the fire over 10 days as it burned through some 40,000 acres. On the night of August 24, strong east winds spread the blaze over 240,000 more acres in only 20 hours, making it one of the fastest-growing forest fires of the 20th century.

Though its spread was eventually stopped by rain, the devastation caused by the blaze primed the region for future forest fires. In 1939, another fire raged over more than 200,000 acres of the Coast Range, including 19,000 acres of previously untouched forest. In 1945, two fires burned 182,000 acres, and in 1951, another two fires consumed more than 32,000 more. All told, the fires of the Tillamook Burn damaged or destroyed a combined total of 355,000 acres (554 square miles) of the country’s richest timberland.

In the years after 1951, much of the land in the Coast Range began moving from private to public ownership, as struggling landholders forfeited their property to the government rather than pay property taxes on the damaged land. With the land under state control, the legacy of the Tillamook Burn continued to shape life in the region for decades to come, as the Oregon Department of Forestry launched comprehensive fire-protection and reforestation programs, including the planting millions of seedlings by hand and via helicopter.

Source: History.com

Detecting and Confirming the Presence of Road Flare Residue in Fire Investigations

Scott Nesvold, M.S., M.Eng., P.E.
Crane Engineering Building Science, USA
Kerri Pacholke, MFS, F-ABC
Crane Engineering, USA

Presented at International Symposium on Fire Investigation, 2012


In a recent fire investigation, a vehicle owner claimed an accidental fire had destroyed his vehicle. An observant fire investigator, who suspected arson based on the facts of the case, found a small pile of “white residue” and some other parts in the debris and sent the residue to a lab to be analyzed. Fourier Transform Infrared Spectroscopy (FT-IR) was conducted as a presumptive test and revealed the presence of Strontium, a common chemical found in road flares. The presence of a high concentration of Strontium was confirmed using a Scanning Electron Microscope (SEM) combined with an Energy Dispersive X-ray Spectroscopy (EDS) unit. Based on the evidence and this analysis, the residue was confirmed to be from a road flare. The vehicle fire was determined to be incendiary based on this analysis.

A test burn was conducted in a furnished condemned house. The house was burned room by room for fire investigator training. One fire was ignited in a furnished living room with a red road flare placed at the leg of a couch. The room was allowed to burn post-flashover for several minutes. Normal suppression and overhaul was performed by the firefighters on the scene. On the following day, a team of fire investigators was asked to determine the source of ignition. The flare residue remained and was visible in the area of origin, but was not discovered or identified by the team of investigators assigned to that particular room.

The testing required to identify flare residue is not included during standard ignitable liquid residue (ILR) tests. Traditional analysis performed on fire debris is for the presence of ILRs which are organic and volatile. Road flare residues are solid and inorganic and therefore, are not detectible using these standard examination techniques.

The residue that remains after burning a road flare is a whitish-grey solidified pool. The color and texture of the white material blends well with gypsum wallboard or plaster fragments typically found after post-flashover fires or fire department overhaul procedures and is therefore easily overlooked. Other components of a road flare may also be present including a cap, wooden plug, metal nail, wire legs, a base or possibly the remains of the cardboard tube.

Road flares are widely available for purchase, and are often included in a typical roadside safety kit. This widespread availability, high burn temperature (1450 °C, 2650 °F) and high heat release rate lends itself as a ready ignition source for incendiary fires. Due to the extended burn times of some road flares, they can be used to delay the start of an incendiary fire which may allow an alibi to be established.

Historically, minimal research has been performed on the role of road flares in incendiary fires. This research investigates the chemical signatures present following a fire that positively identifies the presence of road flare residue. It will also evaluate the remaining components and residue and visual burn patterns that occur when road flares are placed in proximity to common construction materials (such as gypsum wall board, carpet, plywood subfloors, etc.). Finally, it examines the remaining components and residue following vehicle fires.

NFPA 921 requires that the source of ignition, first fuel and the circumstances or conditions which brought them together be identified. The purpose of this research is to assist fire investigators in identifying the possible remains of a road flare during a fire investigation and explain the methods used to confirm the presence of a road flare through FT-IR and SEM-EDS analysis.

Download the complete paper

Fire Effects on High Efficiency Compact Florescent Lighting

Richard J. Meier, CFEI, CFII, CVFI
Staff Fire and Explosion Analyst
John A. Kennedy and Associates
Fire and Explosion Analysis Experts, USA

Presented at International Symposium on Fire Investigation, 2012


The Energy Independence and Security Act of 2007 has mandated that most of the incandescent lights currently in use will be phased out by 2014 and replaced with more efficient means of producing light. Many manufacturers have begun producing compact fluorescent and LED lighting to replace the incandescent bulb. While this is a boon for energy conservation, what will it mean for the fire investigator? For years investigators have used heat distorted light bulbs to help determine the origin and intensity of fires. The purpose of this study is to establish a base of information on the effects of fire on new styles of lighting, and how the effects of fire can aid the investigator in his or her work.

Download the complete paper here

Wind Turbine Fire Origin Investigation

Timothy L. Morse, Ph.D., P.E.
Robert W. Whittlesey, Ph.D., CFEI
Exponent, Inc.

Presented at International Symposium on Fire Investigation, 2014

Wind turbines and wind farms have become increasingly widespread in the United States. Due to the combination of potential ignition sources (electrical failure, overheating of rotating components, lightning strikes) and multiple fuel loads (fiberglass, bearing grease, gearbox oil, hydraulic oil) wind turbine fires are a regular occurrence. Since wind turbine fires often occur in the nacelle, which can be 200 feet or more above the ground, firefighting options are limited. Wind turbine firefighting efforts are usually directed at preventing the spread of the fire to adjacent land or structures, such as by falling flaming debris, rather than extinguishing the fire. As a result, wind turbine fires often burn until the fuel uptower is exhausted and the fire self-extinguishes. This can present a challenge to a fire origin investigation. Many fire patterns that are observed in a nacelle can provide misleading or conflicting information as they may indicate a fuel load or a source of ventilation, rather than the fire origin. Therefore, attempting to use fire patterns alone to identify the origin is often unsuccessful.

A wind turbine fire origin investigation can be greatly assisted by the large amount of data that is recorded regarding the operation of a wind turbine. Wind turbines are heavily instrumented, with sensors throughout the turbine. Position sensors monitor the blade pitch position, the nacelle yaw position, and the rotational speed of the high speed and low speed shafts. Temperature sensors monitor the gearbox oil temperature, the hydraulic oil temperature, and the brake temperature (as well as other temperatures). The performance of the electrical systems (generator, transformer, inverter) are carefully monitored. Wind turbines also often have vibration sensors in various locations. The data from these sensors are used to control the operation of the turbine through the supervisory control and data acquisition (SCADA) system. This system logs the states of all these sensors as often as once every second and records any alarm states.

A detailed review of this logged data can provide essential guidance to the wind turbine fire investigator. The logged data can indicate which systems or components are having problems prior to the fire, identify any rapid changes in operational state proximate to the time of the fire, or show which systems or components are still functional during the fire and when they lose functionality. A full understanding of the location of the different sensors, and where their communication lines run may also provide indications of the direction of fire spread. Any fire origin that is considered must be consistent with the timing and nature of the SCADA data and SCADA alarms.

Download the complete paper

Port Chicago Disaster, July 17, 1944

An ammunition ship explodes while being loaded in Port Chicago, California, killing 332 people on this day in 1944. The United States’ World War II military campaign in the Pacific was in full swing at the time. Poor procedures and lack of training led to the disaster.

Port Chicago, about 30 miles north of San Francisco, was developed into a munitions facility when the Naval Ammunition Depot at Mare Island, California, could not fully supply the war effort. By the summer of 1944, expansion of the Port Chicago facility allowed for loading two ships at once around the clock. The Navy units assigned to the dangerous loading operations were generally segregated African-American units. For the most part, these men had not been trained in handling munitions. Additionally, safety standards were forgotten in the rush to keep up frenetic loading schedules.

On the evening of July 17, the SS Quinault Victory and SS E.A. Bryan, two merchant ships, were being loaded. The holds were being packed with 4,600 tons of explosives–bombs, depth charges and ammunition. Another 400 tons of explosives were nearby on rail cars. Approximately 320 workers were on or near the pier when, at 10:18 p.m., a series of massive explosions over several seconds destroyed everything and everyone in the vicinity. The blasts were felt as far away as Nevada and the resulting damage extended as far as San Francisco. Every building in Port Chicago was damaged and people were literally knocked off their feet. Smoke and fire extended nearly two miles into the air. The pilot of a plane flying at 9,000 feet in the area claimed that metal chunks from the explosion flew past him.

Source: History.com


Optimization of Carbon Monoxide Detector Layout in Residential Structures

Derek Engel, Scott Davis
GexCon US, 8433 Rugby Ave. Suite 100, Bethesda, MD 20814

Presented at International Symposium on Fire Investigation, 2012

The current NFPA 720 code requirement for carbon monoxide (CO) detectors in residential structures requires placement outside of each separate sleeping area and on each floor of the residence. There is however no further guidance to specific placement of the detector (high, low, near or within furnace closets, etc.), as well as no acknowledgement to different housing and HAVC styles (forced hot air, hot water, etc.). As the concentration of CO approaches several hundred parts per million, the time for detector alarm can be as little as a few minutes, much smaller than the characteristic mixing time of the residence. The general basis for detector placement requirements assumes that once the flue gases cool CO is generally neutrally buoyant in air, and becomes well mixed and distributed evenly throughout the residence. Previous investigations have concluded that the CO is well mixed for residences with forced hot air heating systems and the CO in hot flue gases stratifies due to buoyancy for systems without an air-handling device to cause mixing.

Using the CFD software FLACS, a study was performed to evaluate how CO would disperse and migrate in various residential structures and various HVAC designs. The goal would be to evaluate the migration of CO originating from hot flue gases, which are improperly vented into structures, and assess the validity of the well-mixed assumption as well as study the general dispersion patterns. In addition, the study will provide further guidance as to optimal places for detector placement to allow early detection, while minimizing nuisance alarms.

Download the complete paper

Did Nero play the fiddle while Rome burned?

AD 64, Jul 18th:  Nero’s Rome burns

The great fire of Rome breaks out and destroys much of the city on this day in the year 64.  Despite the well-known stories, there is no evidence that the Roman emperor, Nero, either started the fire or played the fiddle while it burned.  Still, he did use the disaster to further his political agenda.

The fire began in the slums of a district south of the legendary Palatine Hill.  The area’s homes burned very quickly and the fire spread north, fueled by high winds.  During the chaos of the fire, there were reports of heavy looting.  The fire ended up raging out of control for nearly three days.  Three of Rome’s 14 districts were completely wiped out; only four were untouched by the tremendous conflagration.  Hundreds of people died in the fire and many thousands were left homeless.

Although popular legend holds that Emperor Nero fiddled while the city burned, this account is wrong on several accounts.  First, the fiddle did not even exist at the time. Instead, Nero was well known for his talent on the lyre; he often composed his own music.  More importantly, Nero was actually 35 miles away in Antium when the fire broke out.  In fact, he let his palace be used as a shelter.

Legend has long blamed Nero for a couple of reasons.  Nero did not like the aesthetics of the city and used the devastation of the fire in order to change much of it and institute new building codes throughout the city.  Nero also used the fire to clamp down on the growing influence of Christians in Rome.  He arrested, tortured and executed hundreds of Christians on the pretext that they had something to do with the fire.

Source: History.com 

Ignition Propensity of Cannabis Cigarettes

Zachary J. Jason, PE, CFEI
Dennis E. Shelp, MS, PE, CFI, CFEI
John L. Schumacher, MChE, PE, CFI, CFPS
Todd J. Hedglin, CFI, CFEI
AEI Corporation
Littleton, CO, USA

Presented at International Symposium on Fire Investigation, 2014


It is well known that cigarettes are the leading cause of fire deaths in the United States.  The National Fire Protection Association (NFPA) reports that in 2011 alone, there were over 90,000 smoking-related fires, contributing to over 540 civilian deaths, 1,640 civilian injuries, and $621 million in direct property damage. However, the NFPA statistics, collected from The National Fire Incident Reporting System (NFIRS) and the NFPA annual survey, define “Smoking Materials” as lighted tobacco products (typically tobacco cigarettes). There is little to no data regarding fires caused by cannabis, or what will hereafter be referred to as marijuana cigarettes.

With the recent legalization of marijuana in the states of Colorado and Washington, pending potential legalization in 13 other states, and 20 states with medical-marijuana systems already in place the availability and usage of marijuana is becoming more commonplace. This raises many interesting questions with regard to fire safety as it relates to marijuana cigarettes. For example, what are the burn times and smoldering capability for marijuana cigarettes? How do marijuana cigarettes compare with tobacco cigarettes in their ability to initiate smoldering combustion in upholstered furniture and mattresses? To date, research regarding these questions has been difficult due to the illegal status of cannabis, and currently very little is known about the ignition propensity and combustion characteristics of marijuana cigarettes. Given the recent changes in Colorado law, however, AEI Corporation has performed some of the first scientific testing of its kind looking at the smoldering and burning behavior of marijuana cigarettes.

This paper outlines the first phase of our research into the overall fire hazards of marijuana cigarettes and compares the ignition characteristics of marijuana to those of tobacco, when tested in accordance with current test methods adopted for the tobacco industry. More specifically, our testing quantifies the ignition strength of marijuana cigarettes and their propensity to ignite soft furnishings based on the parameters set forth in American Society of Testing and Materials (ASTM) Standard E2187-2009, Standard Test Method for Measuring the Ignition Strength of Cigarettes.  The results of our tests evaluating ignition propensity of marijuana cigarettes are presented in comparison to those of tobacco cigarettes tested under the same conditions. In addition, the effects of different variables on the burning, smoldering, and ignition propensity of marijuana cigarettes will be examined.

Download the complete paper

Chicago Fire of 1874

The Chicago Fire of 1874 was a conflagration in Chicago, Illinois, that took place on July 14, 1874. Reports of the extent of the damage vary somewhat, but sources generally agree that the fire burned forty-seven acres just south of the Loop, destroyed 812 structures and killed 20 people. The affected neighborhood had been home to Chicago’s community of Jewish immigrants from Russia and Poland, as well as to a significant population of middle-class African-American families; both ethnic groups were displaced in the aftermath of the fire to other neighborhoods on the city’s West and South Sides.

The fire insurance industry’s National Board of Underwriters responded to the fire by demanding widespread changes in Chicago’s fire prevention and firefighting efforts, and ultimately encouraged fire insurers to cancel all coverage of buildings in the city in October. Many insurers did halt their activities in Chicago, and only returned to issuing policies in the city after the municipal government adopted many of the suggested reforms.

Source: Wikipedia