How ‘Sloppy’ Fire Science Sends Innocents to Prison

The following opinion by Paul Bieber was published by The Crime Report on July 23, 2013.

For George Souliotes, this 4th of July was an Independence Day like no other. It was his first full day of freedom, after 16 years behind bars in California for crimes he did not commit.

Seventeen hundred miles away in Texas, Ed Graf spent his 4th of July in the same manner he has for the past 27 years—an innocent man confined in a state prison. His celebration of freedom will have to wait.

George Souliotes and Ed Graf do not know each other, but they share an unenviable bond: they were both wrongfully convicted of arson and murder and were both sentenced to life without parole.

Souliotes was tried in Modesto, CA for a 1997 blaze in his rental property that killed a mother and her two young children. Graf went to court in Waco, TX for a 1986 fire in his backyard shed that killed his two step-sons. In each case an amateurish fire investigation quickly escalated into a full-fledged witch hunt.

Fire investigators testified with absolute certainty in both cases that deep charring, holes burned into the wooden floors and "pour patterns" could only have been caused by an ignitable liquid, such as gasoline or paint thinner, intentionally used to start the fire. They also testified that each fire burned "abnormally hot,” further suggesting to investigators that the fires were fueled by a liquid accelerant.

The forensic expert testimony in these cases has been shown to be fundamentally unreliable.

The 1992 publication of NFPA 921 Guide for Fire and Explosion Investigation warns investigators not to rely on burn pattern analysis as proof of the presence of an ignitable liquid. A 1997 study by the United States Fire Administration demonstrated that the suspicious burn patterns listed above are created in any fully involved compartment fire, regardless of ignitable liquids, and that the heat of a fully developed fire has nothing to do with the use of a liquid accelerant.

More recently, blind study research by the Arson Research Project has shown the accuracy of experienced fire investigators in determining the presence of a liquid accelerant under these circumstances to be no better than a random guess.

Unreliable burn pattern analysis was so prevalent in the 1980s and 1990s that the Texas Forensic Science Commission recently recommended that all arson convictions in Texas be reviewed in order to determine which of the convictions are based on unreliable forensic evidence. The Texas Innocence Project took up the call and in conjunction with the Texas State Fire Marshall's Office has reviewed hundreds of the state's arson convictions.

The review has narrowed the field to less than a dozen cases, among over a thousand Texas arson convictions, where the presence of unsound forensic fire science is mixed with strong claims of actual innocence.

While fire investigators across the country slowly came to grips with the reality that many previous conclusions might have landed innocent people in prison, the Souliotes and Graf cases moved through the crushingly slow process of appellate review.

With the tenacious efforts of the Northern California Innocence Project and pro bono attorneys, Souliotes’ case found its way through state and federal courts. It landed in an evidentiary hearing in front of U.S. Magistrate Judge Michael Seng, who concluded that “no reasonable jury would have found him guilty beyond a reasonable doubt.” In a confirming opinion, U.S. District Judge Anthony Ishii said that Souliotes had shown “actual innocence.”

Late last month, the independent Texas fire review panel convened by the State Fire Marshal’s Office concluded that the original determination of arson by Texas fire investigators in the case of Ed Graf was mistaken. Speaking of the Graf case, Texas State Fire Marshal Chris Connealy said the original investigators “failed to meet the present day standard of care.”

The Souliotes and Graf cases represent far more than simply not meeting a basic standard of care. They represent a nationwide travesty: accidental fires continue to be misidentified as arson leading to wrongful convictions and at least one execution.

Convicted of arson and murder for the 1991 fire that killed his three young children in Corsicana, TX, Cameron Todd Willingham was executed in 2004. His conviction was based on the same discredited evidence present in the cases of Souliotes and Graf. Review of the Willingham case by the Texas Forensic Science Commission and independent experts has confirmed that the fire leading to his conviction was almost certainly accidental, not arson, and that the forensic determinations of the state’s experts were not based on science.

In the case of Willingham, sloppy fire science, mixed with a good dose of bias and speculation, led to the execution of an innocent man.

In spite of the presence of undisputedly flawed forensic evidence, a complete lack of reliable evidence to support a determination of arson, and strong showings of actual innocence, local prosecutors in California and Texas promised to go forward with retrials in both cases.

After protracted negotiations with Souliotes’ attorneys, on the eve of trial the Stanislaus County District Attorney agreed to drop the charges of arson and murder in exchange for a no-contest plea to involuntary manslaughter for failure to maintain the smoke detectors in the rental property where the fire occurred.

Under the terms of the plea, Souliotes did not admit guilt; but he did acknowledge that prosecutors had sufficient evidence to prove their allegation that the smoke detectors were not properly maintained.

The re-trial of Ed Graf is scheduled to begin in Waco in the coming weeks.

Two things must be done. First, charges should be dropped against Ed Graf and he should be immediately and unconditionally freed. Like George Souliotes, Ed Graf has been in prison for dozens of years for a crime that was never a crime at all.

Second, the same type of arson review conducted in Texas should be performed in arson cases nationally. The Texas Innocence Project and State Fire Marshal’s Office are proactive and forward thinking in their review of Texas’ arson convictions. Convictions outside of Texas deserve no less.

On Wednesday afternoon, July 3rd, George Souliotes walked through the lobby of the Stanislaus County jail towards the front doors that separated him from his first steps of freedom. After hugging his family and defense attorneys, while wiping tears from his eyes, he pointed through the glass of the front door.

“I see the sun,” he said. “It’s beautiful”.

It is time for Ed Graf and others languishing in prison on wrongful arson convictions to see the sun.

Paul Bieber is a private investigator specializing in indigent defense investigation and the founder and director of the Arson Research Project. He welcomes comments from readers.