A recent report by the American Association for the Advancement of Science examines some of the shortcomings of current forensic fire investigation practices, both on the scene and in the lab in cases which may include suspected arson, and offers several suggestions for bridging the gaps that have led to wrongful convictions, and possibly some overlooked cases of intentional arson.

One major finding of the report, published last week, was the unreliability of fire investigators to determine the origin of a fire in rooms where the fire had reached flashover—a point at which extreme temperatures of about 500-600 degrees Celsius (932-1112 degrees Fahrenheit) cause every ignitable surface in a room to immediately burst into flames. The report describes this as the point at which “a fire in a room becomes a room on fire.”

The report mentions two studies, one of which found that burn patterns created by flashover can produce error rates of over 75 percent when trying to determine the origin of the fire. Another study found that only 13 out of 53 fire investigators could correctly determine which quadrant of a room a fire was started in after the room burned post-flashover for only three minutes.

Other problems identified in current fire scene investigation practices included the persistence of outdated, inaccurate methods taken from previous standards and literature that have since been discredited, and the lack of efforts to prevent potential bias in fire investigations, when investigators are involved in managing other areas of the case and don’t receive enough training to avoid cognitive bias when exposed to outside information that could influence their judgement.

Suggestions for improving the state of modern forensic fire investigation included conducting research to study burn patterns, including full-scale and reduced scale tests in realistic conditions; comparing computer-based fire models with the results of physical fire tests; refining the use of canine units at fire scenes, which are considered more efficient than technological counterparts such as electronic “noses”; and reducing the likelihood of bias-tainted judgments by fire experts through cognitive training, “blinded” procedures and separation of scientific investigative roles from case management roles.

The report concluded that lab analysis of debris from fires was not plagued with as many problems as on-the-scene analysis, but that “the chemical evidence found at the scene of a fire can still be confusing and misleading.” The report pointed out that some materials, when burned, can produce chemicals that can be mistaken for ignitable liquid residues (ILRs) often used as accelerants. This can make it difficult to distinguish whether chemicals detected in fire debris are the result of an arsonist’s use of an accelerant or the result of furniture and other materials combusting.

Suggestions for improved forensic fire analysis in the lab included improvement of field chemical testing tools such as mass spectrometers and more sensitive electronics “noses,” to collect data on the scene that could be useful for lab analysists; adherence to current testing standards set forth by the American Society for Testing and Materials (ASTM), which are considered to be scientifically accurate; research into further improving ASTM methods and into quantifying the rates of false positives and false negatives in the testing of fire debris for ILRs; and further research into ILRs, the chemicals that can imitate IRLs and the impact of microbial degradation at fire scenes.

The report mentions the impact that faulty fire science has had, pointing out that 63 individuals have been exonerated of arson convictions since 1991, according to the National Registry of Exonerations. It adds that the number of arsonists who have gotten away with their crimes due to faulty fire science cannot be known.

One of the most infamous cases to raise questions about accuracy in arson investigations was that of Cameron Todd Willingham, a Texas man who was convicted of murder after investigators said the fire that killed his three young children was intentionally set by him. Following his conviction, Willingham was executed in 2004.

Doubts about Willingham’s guilt and the accuracy of the forensic science argued in the case have been widely raised since his conviction and death, although he has never been formerly exonerated. In 2010, a panel of the Texas Forensic Science Commission concluded that “flawed science” had been used in the case, but that the investigators in the case could not be considered negligent, as they were following the accepted standards in place at the time, the Houston Chronicle reported.

The most recent arson exoneree, according to the National Registry of Exonerations, is Adam Gray, who was convicted of first-degree murder and aggravated arson at the age of 14 after two of his neighbors died in a fire at his two-flat apartment building. During his trial, fire experts argued that burn patterns found in the apartment indicated the use of an accelerant, and a lab analyst testified that a “high-boiling petroleum distillate” was found in a milk jug allegedly used to distribute the accelerant. Gray was sentenced to life in prison, but released on May 3, 2017 after two witnesses recanted their testimony, and a review of the case found that the science used to convict him had since been debunked.

The full AAAS report is available here:

A plain language version of the report is available here: