On the New Science and Past Mistakes of Fire Investigation

The following is an excerpt from Burned: A Story of Murder and the Crime That Wasn’t, by Edward Humes. In this impassioned expose of junk science, bad forensics, and other ways innocent people end up in jail, Humes recounts the case of Jo Ann Parks, convicted of murdering her three children in a house fire, only to have her case reopened through the actions of the California Innocence Project 26 years after her conviction. The following passage explores recent breakthroughs in the science of fire investigation that allowed her case to be reexamined, and looks at the uncomfortably slow progress of new ideas that challenge the status quo. 

__________________________________

Three Days in October 1991

The worst fire to scorch the California Bay Area in 140 years began with just a few acres of blackened grass.

Initially, this seemed nothing more than an inconsequential accident on private land. Fire crews snuffed out the little blaze before it could reach the winding residential streets overlooking the steely arc of San Francisco Bay. Five years of drought had left the Oakland Hills area vulnerable to the merest spark, the desiccated grass crunching underfoot with the sound of broken glass. Yet it seemed luck prevailed and disaster had been averted on October 19, 1991.

And that would have been the end of it, but for an unlucky change in the weather: The Diablo winds kicked in on the morning of October 20. Blistering, dry, and moving at more than forty miles an hour, the Bay Area’s analog to Los Angeles’s legendary Santa Ana winds fanned to life small embers hidden beneath that patch of burnt grass. Then the gusts drove the renewed flames up into the maze of hillside homes with their lovely wood siding and oh-so-flammable shake rooftops.

One by one, hillsides, houses, parks, and apartment buildings succumbed to the spreading flames, until the Bay Area had a full-blown wildfire on its hands.

There weren’t enough firefighters, hoses, or hydrants to stop what had become a wall of fire and choking smoke, even as an army of fire engines raced to the scene from dozens of communities, some as far away as Las Vegas. For every house where a fire crew knocked down the flames, five others on the same street would be gone by the time the firefighters were ready to move on. The fire grew so fierce that the resinous pine trees that covered the hillsides didn’t just burn. They exploded.

The Oakland firestorm, as the wildfire came to be known, destroyed more than three thousand homes before the Diablo winds abruptly shifted and died around nine that night. Hundreds of people were injured. Twenty-five burned to death. Ashes rained down on Candlestick Park, packed with football fans twenty miles away. A sooty, greasy stench reminiscent of rancid burnt popcorn blanketed the landscape. Whole neighborhoods had been reduced to rubble, houses burned down to their foundations, leaving only brick chimneys standing. Sagging metal plumbing rose from the wreckage like charred dinosaur fossils. The twisted plumage of piano wire bristled from the black stumps of Baldwins, Steinways, and Yamahas. Gray powder coated everything.

In the aftermath, a crew of arson investigators and fire scientists descended on those spectral flame-scoured neighborhoods. Their mission was unique. They came not to do their usual thing, which is to figure out why a fire happened, and what—or who—was responsible. For once, the investigators already knew the fire’s origin and how it spread: an accident of drought and complacency, and a seemingly extinguished brush fire resurrected by the devil winds. Instead, the team came to this disaster scene to learn about other fires.

It’s not hard to see what attracted them. Here stood home after devastated home, many reduced to what firefighters call “black holes.” Most had burned completely, until there was nothing left to burn, all without human intervention, without any arsonists splashing gasoline or kids playing with matches or, for a majority of homes, firefighters interrupting the process and destroying the evidence with their hoses, axes, foam, and shovels. The team of experts came to see and learn what a known accidental fire looks like when allowed to burn itself out, so they could find ways to distinguish these black holes from true arson fires.

Black holes are among the most difficult fires to investigate. The flames not only destroy buildings, they also scour away almost every bit of evidence of a crime. Investigators had long relied on a few tricks of the trade to find signs of arson in a black hole. Now the disaster in the Oakland Hills provided a real-world laboratory to test, refine, and add to those tricks. The team was drawn by the promise of finding some small silver lining beneath all those tons of soot, ash, and loss.

They found something, all right. Just not what they expected.

They found, to their amazement, signs of arson. Everywhere.

Their response as they walked through the blackened hills provided a curious mirror to Ron Ablott’s reaction to the scene at the Parks apartment. They kept saying to one another, “This isn’t right. Something’s not right here.”

They found crazed glass—windows fractured by webs of tiny cracks. They had been taught, like most arson investigators had been taught for decades, that crazing indicated the sort of fast-moving, very hot fires caused by the use of gasoline or some other flammable accelerant. Crazing was supposed to be a sure sign of arson. Yet they were finding it in one out of every four houses they visited—houses where they knew no arson had occurred. Subsequent experiments showed that crazing occurs not from rapid heating but rapid cooling—such as occurs when fire hoses douse a flaming building.

Even more common were twisted, melted bedsprings in the remains of bedrooms, another one of the tricks of the trade that, for decades, had been used as a sign of arson in a black-hole ruin. Melted copper plumbing and fixtures, and melted metal doorway thresholds turned up everywhere, too, which were only supposed to happen in the rapid, extreme fires caused by an arsonist with a gas can. Spalling of concrete, a surface chipping that was supposed to indicate the use of accelerants, also turned up.

John Lentini, a former police forensics expert working for a private fire investigation consultancy, was on the team of scientists digging around in the Oakland firestorm ruins. He would go on to the role of lead author of the report on the group’s findings, which could be summarized very simply: Indicators of arson that had been passed on across generations of fire investigators, and that had been used to send people to prison for decades, were wrong. The same supposedly suspicious indicators occurred in accidental fires, too.

The Oakland firestorm investigation was undertaken to expand and refine scientific theories about fire and arson. Instead, it had become a real-world laboratory for disproving them. Those theories had never really been tested before in a scientific way. They merely had been the received wisdom in the fire investigation business, their admissibility in court becoming a legal precedent and never questioned after that—the way doctors had been taught for centuries that bloodletting of patients sick with fevers or infections was an effective treatment.

This became a pivotal moment for Lentini. He had been hired just months before to help convict a Jacksonville, Florida, man accused of setting a 1990 fire that killed his pregnant wife, his sister, and his four nieces and nephews. The man claimed his son had accidentally set fire to the living room couch while playing with a cigarette lighter. But the police arson investigators had found V-patterns that suggested a fire had been set elsewhere in the house, and “pour patterns” on the floor that suggested gasoline had been splashed there and set aflame. Dubbed the “Lime Street fire” in subsequent news accounts, the case had seemed an obvious multiple murder by arson, the arrest a foregone conclusion. But there was no seeming motive for the crime, and the prosecutor wanted an airtight case. So Lentini and a colleague were hired to see if there was any possibility that the fire could have started accidentally as the defendant claimed, or if the evidence proved this fire could not possibly have begun with a child, a couch, and a lighter.

It turned out there was an identical house nearby that was slated to be torn down. Lentini and his colleague decided to carpet and furnish this home in the same way as the Lime Street house where the fatal fire occurred, install video cameras, and set fire to the couch just as the suspect had claimed. Firefighters would stand by to extinguish the flames at the same stage at which the original fire was put out. Lentini expected the results would reveal the accident story as an impossible lie, and the prosecutor could proceed with putting the man away—possibly on death row—with righteous confidence.

Instead, the same suspicious burn patterns were found in the test house after the fire was put out. The indicators of arson had appeared where they should not have appeared.

The culprit: flashover, the transformation of a fire in a room into a room on fire. Flashover had occurred in both the fatal fire and in the Lime Street re-creation, and it was this then-poorly understood phenomenon that revealed serious problems with the traditional arson investigation playbook. Normally, the largest V-patterns and the area of greatest damage in a burning structure can reliably indicate where a fire started. The logic of this is obvious: The most damage generally occurs where the fire burns the longest, and that would be the starting point in many simple structure fires. But this rule of thumb doesn’t work so reliably after flashover, when large V-patterns can form in areas away from the fire’s starting point, and burn marks that mimic flammable-liquid pour patterns can emerge as well. When Lentini reported his findings, the prosecutor, instead of shoring up his case, ended up dismissing it. And an innocent man was saved.

Had the prosecutor not requested that test, the man who lost his wife, sister, and nieces and nephews almost certainly would have ended up convicted of a crime that never happened.

“I had come within twenty-four hours of giving testimony that could well have sent an innocent person to Florida’s electric chair,” Lentini would later say. “Needless to say, I was chastened by the experience.”

He was more than chastened. The combination of the Lime Street fire and the Oakland firestorm changed the course of John Lentini’s life. He realized that key elements of fire science as practiced in 1991, and for decades before, were wrong. Fatally wrong. There was no science to back up those beliefs. And hardly anyone even knew there was a problem.

But Lentini knew lives had been ruined because of this. Insurance payoffs had been denied, leaving families bankrupt and homeless. People had been sent to prison. He had sent people to prison—for crimes that might never have happened. Not every case, he knew. Probably not even a majority of arson cases, most of which were so simple or so obvious that such glaring injustices were all but impossible. But there would be enough that were wrong, that were infected by flawed science. Too many. Thousands of criminal and civil actions had been taken against people based on what amounted to fire mythology.

Lentini walked away from Oakland a shaken man, but a man with a mission: to put the science into “fire science” where it should have been all along. He would volunteer to review for free cases in which bad arson theories might have led to wrongful convictions. And he would lobby for national guidelines for how a proper fire investigation should be conducted.

A year later a committee of leading fire experts, Lentini among them, was organized by the National Fire Protection Association. This group published a landmark report with the bland title of NFPA 921. It set guidelines for a proper fire investigation intended for use nationwide by every fire department and in the investigation of every fire.

One of the key recommendations: Fire experts, including field investigators for police and fire departments, should follow the scientific method when rendering an opinion on the origin and cause of a fire. If, for instance, window crazing was present and introduced as evidence, an expert should have to prove with real science and real-world data that only an arson fire could cause crazing. That had never been done before and, after Oakland, never could be done, at least for crazing. All the other long-held beliefs about the indicators of arson would have to be backed up with clear scientific evidence from now on to comply with these new guidelines.

The NFPA 921 recommendation—which basically requires testing, data, and proof before a scientific finding can be accepted as grounds for sending a person to prison—sounds innocuous and commonsensical. But it created, appropriately enough, a firestorm.

Firefighters, law-enforcement organizations, and the National District Attorneys Association fiercely opposed the new guidelines, which are influential but entirely voluntary. Expert opinions long considered rock solid, backed by legal precedent, and used to put away accused arsonists for years would be undermined. Old cases would be opened. New prosecutions made more complex and difficult. Worst of all, accomplished investigators who were just doing what they had been taught to do, who were acting in good faith even as they practiced bad science, would have to admit in an unknown number of cases that they had screwed up—and that those screwups had come at a terrible price.

Most, at least at first, found it easier to dispute rather than embrace the new guidelines, calling them misguided and mistaken. Some argued that they were the handiwork of the defense bar trying to get guilty people off the hook and out of jail. Others saw in the new guidelines the well-meaning naiveté of eggheads in spotless lab coats who practiced science in clean laboratories, but who lacked practical experience with the dirty business of crawling through ravaged, ash-filled fire scenes in search of truth.

One of the leading law-enforcement arson investigators in Los Angeles testified in April 1992 about his colleagues’ unfavorable attitude toward the then-new NFPA 921. He dismissed the National Fire Protection Association, which produced the guidelines, as a group primarily interested in promoting better sprinkler and fire-door designs. And he opined that the group’s investigation guidelines conflicted with techniques embraced by most arson investigators “worth their salt.”

That leading investigator was Ron Ablott of the Los Angeles County Sheriff’s Department Arson/Explosives Detail—the principal fire investigator in the Jo Ann Parks case.

__________________________________

Excerpted from BURNED: A STORY OF MURDER AND THE CRIME THAT WASN’T by Edward Humes. Copyright © 2019 by Edward Humes. Excerpted by permission of Dutton Books, a division of Penguin Random House LLC. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.