Around 4:00 p.m. on June 30, 2013, a 30-year-old hotshot named Christopher MacKenzie pulls a camera from his pocket and shoots a short video. Downhill from him are ten firefighters, all members of the Granite Mountain Hotshots. In the background, the Yarnell Hill Fire sweeps toward the 650-person town of Yarnell, Arizona. On the radio in the background is the voice of the crew’s superintendent, Eric Marsh. He’s somewhere nearby and hints at playing witness to coming disaster. “I was just saying I knew this was coming,” he says. “When I called and asked you what your comfort level was, I could just feel it—you know, too bad.”
“I copy,” says Jesse Steed, the crew’s second in command. “And it’s almost made it to that road we walked in on.”
In that moment, the fire was exploding with a fury most of the several hundred firefighters battling the blaze remembered only as “unprecedented.” One called it “pandemonium.” MacKenzie and his crew were watching this unfold from the safest place on the fire: in the already burned brush high on the same ridge where lightning had started the blaze two days earlier. Less than 50 minutes later, MacKenzie, Steed, Marsh, who had rejoined the crew, and 16 other hotshots were dead in a canyon a mile and a half away, burned to death a short walk from the safety of a ranch on the edge of Yarnell.
The death of 19 Granite Mountain Hotshots, which I wrote about for this magazine and, later, in a book, marked the worst wildland fire disaster in almost 100 years. In the hours after, the Arizona State Forestry Division commissioned a report to find out what happened. Why had the men left the safety of the ridge? For three months, a team of 18 interagency investigators combed over any shred of evidence they could find. They interviewed every firefighter of consequence working the blaze. They took Granite Mountain’s sole surviving member, Brendan “Donut” McDonough, back to the knoll where he last saw his crew. They scoured dispatch records, weather and fuel data, photos, social media posts from firefighters on the blaze, and accounts from civilians. MacKenzie’s partially melted camera was found, having survived a fire that burned hotter than 2,000 degrees. All of the information the investigators collected went into a 116-page record of the tragedy that they hoped could be studied to avoid similar incidents. Yet the investigation felt incomplete. After MacKenzie’s video, the record went spotty for the critical window between when the hotshots left the safety of the ridge and when they reappeared in the canyon minutes before their deaths. Nobody can say for certain why they left.
“All of a sudden, all this other chaos happened. The clarity, the certainty,” says Brad Mayhew, tossing his hands up like he’s throwing confetti. He served as the lead investigator on the report commissioned by the state forestry division immediately after the fatalities. This fall, he agreed to meet me in Yarnell and walk the hotshots’ final steps. It was late afternoon and 100 degrees on September 11. We sat where MacKenzie had shot the video, looking out at the long valley. In the years after the fire, the valley has regrown green but is not yet shaggy. Surrounding us were pyramids of small rocks stacked atop bigger boulders. Mayhew, who is 38, with a salted black beard and a voice that’s deep like that of James Earl Jones, pulled up MacKenzie’s video on his phone to confirm our location. It immediately became clear that somebody had piled the rocks to mark where the ten hotshots had sat or stood in MacKenzie’s final video. “It’s somber,” he mustered.
For most of an hour, we sat among those stones, eating nuts while talking with a big view of the landscape where the tragic fire burned. All the unknowns surrounding Granite Mountain’s deaths bred distrust and blame, and all that emotion soon translated into lawsuits from the some the hotshots’ families. Collectively, they sued the state for wrongful death, settling for $670,000, which was divided among the aggrieved. The fire community clammed up after the lawsuits. “Just talking about Yarnell became radioactive,” Mayhew says. A warm wind was pulling up from the desert and blowing across our backs. “How can this profession make progress if people aren’t comfortable talking about it publicly?”
Most fire fatalities have forced significant safety and cultural changes to wildland firefighting. In an age when fires are getting more dangerous and the need to fight them more pressing, what, if anything, has changed after Yarnell?
In January 2014, 11 veteran firefighters from the nation’s biggest fire agencies—the vanguard of fire, as they were described to me—met in Yarnell. They hiked along the route the hotshots had likely taken from the ridge into the canyon where the 19 died seven months earlier. They arrived at a startling conclusion. “We could see ourselves making the same decision they’d made,” said Travis Dotson, a member of the Wildland Fire Lessons Learned Center, a federally funded organization that helps firefighters improve their performance. Around the time of the field trip, Dotson and others formed an underground group called Honor the Fallen. Included in its couple dozen members were some of the highest-ranking firefighters from the various agencies in the wildland fire business: the Forest Service, Bureau of Land Management, and the Park Service. Their goal was to make sure Yarnell Hill, the most publicized event in wildland firefighting history, forced some much-needed changes to the job’s outdated culture. Three years later, they tried to spark “an age of enlightenment” in wildland fire. As Dotson distilled the shift in mindset, “Before Yarnell, it was about getting better at fighting fire. After, it’s been about getting better at accepting death.”
Some context is needed here. Since 1910, more than 1,100 wildland firefighters have died on the line. “There has never been a fire season that we’ve escaped with no deaths, and many years reach well into the double digits,” says Dotson, who used to be a smokejumper. “Making it through a fire season without a death is a statistical impossibility.” Historically, fire agencies responded to fatalities with investigations that sought to understand what happened. Since 1990, when a blaze killed six firefighters on an inmate crew, those investigations seemed intent on proving that dead firefighters broke rules—sometimes in ways that were criminal. This fervor peaked in 2001 in Washington state, when a fire killed four and the incident commander was charged with involuntary manslaughter. Traditionally, the agencies used the investigators’ conclusions to develop new learning tools, scientific labs, and, mostly, rules. Fatality fires spawned the “18 Watch-Out Situations,” the “Ten Standard Firefighting Orders,” and the ever-growing 118-page Incident Response Pocket Guide that most firefighters keep in their pockets today. Need a reminder on unexploded ordnance safety? That’s on page 27. A refresher on the alignments of patterns for dangerous fire behavior? Page 73. Best practices for a media interview? 111. It’s an astonishing document that matches problems to solutions, but it’s also something like the pamphlet a scout leader might hand a Boy Scout before dropping them into the Alaskan bush. Over time, the relationship between tragedy and rulemaking sewed into the culture the belief that firefighters die only when they break rules.
From the outset, the members of Honor the Fallen understood that Yarnell was unlikely to result in any official change. For one thing, Mayhew’s investigation was of a new wave that borrowed from the military’s tradition: They tried to understand what the firefighters knew in the moment rather than seeking fault in behavior. Instead of chasing “the instant gratification of new rules,” as Mayhew put it, they put the onus of making change on the fire agencies at large. But the approach seemed to fall flat. Granite Mountain was the rare unit operated by a municipality, and the big wildland firefighting agencies did all they could to publicly distance themselves from a tragedy that wasn’t their own. “We treated this whole thing different because Granite Mountain had a different color blood,” Dotson says.
Yarnell did prompt a modest update to the fire shelter, the flimsy aluminum heat shields the hotshots had died under, and the development of a new phone app that helps firefighters get weather updates in real time. But as Honor the Fallen predicted, it led to no significant policy changes.
Then 2015 happened. That year, more acres burned than at any point in recorded history, and the Forest Service lost seven firefighters. That agency is one of many in today’s ballooning wildland fire business, but as the oldest and largest, it sets the industry’s culture. The chief at the time, Tom Tidwell, responded as tradition dictated. “He said, ‘I don’t want another fire season like 2015,’” says John Phipps, director of the Forest Service’s Rocky Mountain Research Station. Tidwell had called him at home in Colorado late one November night and said, “I’m directing you and the leadership team of the Forest Service to come up with a way that we don’t have that kind of a season ever again.” He gave Phipps’s team six months to come up with a way to stop firefighters from dying on the line. “We have 10,000 firefighters,” Phipps remembers thinking. “Well, gee, what can we do in five to six months, get it deployed, and have it make a difference so that everybody goes home in 2016?”
They called it the Life First Initiative. It focused on “reducing the amount of unnecessary risk” to firefighters’ lives. Tidwell’s directive reinforced that the Forest Service “accepted no loss of life” and suggested 11 more rules. (A couple examples: “Under no circumstances will mop-up be allowed under snags or fire-weakened trees.” “Firefighters are prohibited from working alone without radio communications or easy access to emergency medical skills.”) It provided firefighters no tools to assess risk or determine how much of it was necessary.
Because of swift internal backlash, these rules fell short of implementation, but they set the initiative’s tone. From the moment Life First came out, Honor the Fallen considered it a relic. The initiative didn’t mention that over the past three decades, the Forest Service’s fire force had mushroomed from 6,000 employees in 1998—about a third of the agency’s workforce—into a seasonal army that now gobbled up half the agency’s $4 billion–plus annual budget and then spent hundreds of millions more in emergency funding. It didn’t mention that wildland firefighters’ primary job was no longer to save publicly owned trees for the timber industry to cut, but to place themselves between watersheds, infrastructure, cities, and often uncontrollable fires like Yarnell Hill. And most damning, it failed to acknowledge what the agency’s scientific arm openly states: that because of climate change, sick forests, and explosive population growth, every trend points to firefighters being asked to take bigger risks more often. The year after Life First’s release, 15 more firefighters died.
Honor the Fallen responded immediately after Life First’s release. “They attempted to deal with increased complexity with more rules, some of which just show a total disconnect from the reality of today’s wildland fire environment,” says Mark Smith, a consultant for Mission-Centered Solutions, a company with a 20-year history of advising the Forest Service on leadership and culture. “If you accept that zero fatalities is unachievable, why would you establish it as an objective?” On behalf of Honor the Fallen, Smith penned an essay called “The Big Lie,” in which he slaughtered the sacred cow that we could fight fires without firefighters dying. It was time to move the profession out the 1970s and into the 21st century, he argued. In other words, it was time for management to ignore the politics and accept that fighting wildland fire is a dangerous profession.
According to Smith’s calculations, at the start of each season, every wildland firefighter has a one in 1,600 chance of ending up in a coffin by year’s end—and that doesn’t factor in serious injuries or near misses. With an average of 19 deaths a year, the job is roughly as dangerous as a soldier in training, a career where recruits sign a will when they walk into boot camp. (Because of Honor the Fallen’s work, some crews now ask their firefighters to do this.) The strange thing, Smith says, is that the Forest Service’s official policies still insist that more rules, or following existing rules better, would keep everybody alive.
Smith calls this paradigm a “lawyer’s dream,” where the agency has unintentionally created “a cover your ass” environment by requiring that its firefighters follow rules that simply cannot all be simultaneously followed. While these rules are well intentioned and do indeed save lives, he says they also impose a false sense of control in a wildly chaotic environment. My favorite line of Smith’s from “The Big Lie” is this: “There is nothing low risk about a 19-year-old hotshot driving an ATV loaded with fuel mix down a burning mountain at dusk after working a 12-hour day.”
Using formulas developed in the military, Smith, a former Army Ranger, calculated that the vast majority of firefighting operations exist in the medium- to high-risk zone. In other words, there’s a relatively high probability that a tree eventually crushes you, you step on a bee nest, grab the business end of a chainsaw, or get burned. Yet somehow, most firefighters Smith polled believe they work in a low-risk environment—something more like a factory floor. He says that in the special forces, if the Rangers found it too dangerous to take an objective, they came up with a new plan. That’s not always the case in wildland fire.
“It doesn’t matter if it’s one house or one community,” Phipps says. “It’s not part of our protocol to say, gee, we’ll risk less here because it’s only one house.” Put another way, under the current paradigm, the agency regularly risks the same number of firefighters’ lives to save an outhouse as they do the city of Denver.
When I asked Smith how a job as obviously dangerous as wildland firefighting came to be seen as safe, he reached back to 1910, when the Big Burn ripped a 3 million–acre hole into rich timber lands, killed approximately 76 firefighters, and kicked off the Forest Service’s 100-year transition from a land management agency to one of the world’s largest fire departments. He says that back then, the firefighters were militias of men rousted from bars or ranches, and the public wasn’t all that concerned when they died. But die they did, and in great numbers: 25 in California near Griffith Park in 1933; 15 near Cody, Wyoming, in 1937; 11 in the Cleveland National Forest in 1943. In some ways, not much has changed. “Until now, there’s been this insidious cultural legacy where the belief has been if we can just get these low-paid resources to follow the rules, nothing bad will happen,” Smith says.
The backbone of the fire service remains young men and women. Wildland firefighting is a seasonal job with a starting base pay of about $1,920 per month. An Army private makes slightly more than that, and their meals are paid for, plus all their lodging, retirement benefits, and 100 percent of their dental, medical, and vision insurance. Basic training for a soldier is three months. A rookie firefighter can battle blazes as intense as Yarnell if they can pass a week’s worth of online classes and heft a 45-pound pack over three miles in less than 45 minutes.
One reason young men and women might embrace the risk of firefighting is that the job promises big adventure. At least that was true for me when I was in my early twenties and fought fire. But it’s also true that slim budgets and great societal expectations drive risk onto naive kids. The entire Forest Service’s budget is a fingernail on the arm of the military’s—$4.7 billion versus $717 billion. Yet every time a fire starts in a town’s backyard, politicians and the public demand an immediate and forceful response. Smith’s worry is that if the Forest Service admitted the incredibly high chance of death their people are exposed to, their firefighters—or maybe their families—might demand fair compensation. And what land management agency can afford to pay that?
After writing “The Big Lie,” Smith followed up with another piece called “When Luck Runs Out.” In it, he argued that not measuring risk or reward is completely at odds with the military (including the Coast Guard), commercial diving, or almost any other high-risk industry where accidents are accepted as an inevitability. He developed a chart that explained how wildfire agencies might adopt the technique. At the bottom, recreation lands and roads justified a low level of risk. Domestic animals and critical watersheds justified a medium risk. Smith felt a high risk was acceptable if they were working to mitigate threats to regional employment centers and human life. And undertaking extreme risk was OK in only one case: “viable and saveable human life in imminent danger.” It’s widely assumed that describes Granite Mountain’s intent on Yarnell Hill.
This type of risk assessment isn’t yet being done on the fireline. “I see these changes taking ten years, maybe 15,” Smith says.
But there’s reason to be encouraged. Independent of Life First, the Forest Service’s research arm is developing new ways to assess risk. Think of it as the Moneyball of firefighting. The project is led by Dave Calkin, an economist and numbers geek who works for the Forest Service in Missoula, Montana. Calkin is controversial in the fire service. His previous work has shown that the best tactic to take with fires burning under extreme conditions, like California’s 230,000-acre Carr Fire that spun up a tornado of flames and killed six people last July, is to treat them like a hurricane and get the hell out of the way. Lately, Calkin’s been applying economics to weigh the potential of tragic outcomes against the values firefighters try to protect. Quantifying these variables, he says, is the future of wildland firefighting. To sum up his work, Calkin quotes another Forest Service rule about when to fight a wildfire: “The right place, the right time, the right reason. Up until now, the right reason has been left to firefighters to determine,” Calkin says. “That should be a decision made by leadership.”
Ideally, his work will help leadership decide when firefighters should be sent in and when they should wait. Models by Calkin and his team rely on layers of overlaid data. His maps show roads, ridges, rivers—all the typical things found on a map. But they also show vegetation types (forest, brush, the density of dead trees compared to live ones), the perimeters of historic wildfires, and any perceived value at risk—owls, watersheds, towns. His team inputs current and forecasted weather for any given fire. The computer then determines the characteristics of the places where historic fires have stopped and where they haven’t and translates that information into a sort of paint-by-numbers risk map: red where a fire’s most likely to be most dangerous, green where firefighters have the best chance of stopping it, and yellow where they don’t. His models use computers to scout fires and, by doing this, help remove emotion from risk assessment. So far, Calkin hasn’t run a simulation on the Yarnell Hill Fire—there’s no need to since the fire has already burned. But had they run the model on June 30, 2013, it almost certainly would have computed the risk as extreme and the likelihood of success at low to none.
“When we commit firefighters, we want to make sure that the value a firefighter is protecting is worth the investment of the risk they’re exposed to,” Calkin says. That’s not happening, yet. Currently, big agencies fight single fires for months on end, and the public seems content to fund the effort. But nobody is asking if it’s working. That’s because it’s hard to quantify the impact firefighters have on fires. How much bigger would California’s 460,000-acre Mendocino Complex, now the biggest fire in state history, be if $100 million hadn’t been spent trying to control it? Would more than 9,000 homes have burned in last year’s Sonoma and Napa Valley fires if 11,000 firefighters hadn’t tried to stop them? Would they have killed fewer than 42 people? Calkins says too often, regardless of how the fire’s behaving, the assumption is yes. And that means big agencies keep shuttling hordes of firefighters toward the flames without knowing if they can actually do anything to stop them.
Last summer, Calkin’s tool was first put to the test in Arizona’s Tonto National Forest, where the Forest Service let a wildfire burn outside the small town of Globe based on his model’s predictions. It provided recommendations on where they could catch it should they need to, thereby allowing the agency to actively manage the fire while not necessarily fighting it. In the end, it burned 9,000 acres of fire-adapted forest, restoring health to the woods while thinning out some of the excess vegetation that may have otherwise put Globe and the firefighters sent in to protect it at greater risk should a fire spark on some dry and windy day in the future. That project was a dust fleck on the lens of forested lands that need restoration, but it represents a completely different approach to risk mitigation: one that prioritizes maintenance and calculated risk over a reactionary policy of total suppression.
“We’re trying to change a proud tradition,” says Chris Dunn, Calkin’s colleague, who works at Oregon State University. “What we ultimately want to do is help firefighters become fire stewards.”
Back on the ridgetop, Mayhew plays the video MacKenzie shot here five years ago. There’s a moment where the video jumps that looks like an edit. “People seized on that and said we’d doctored the clip,” says Mayhew, shaking his head. “They discounted the entire investigation because they thought they’d caught us in a lie.” In fact, it was two separate but complete clips edited into one. Many firefighters don’t trust investigations. History gave them good reasons. “That’s because for a long time they went out and created reasons to blame workers,” Mayhew says. As an independent contractor, he has made investigating fireline accidents his career.
The team’s reaction to Mayhew’s investigation was particularly strong. He thinks that’s because their investigation did what few others have before. They acknowledged that firefighting is high risk and people sometimes die doing it. In the final report, they didn’t cast blame, which made it harder to learn from the deaths and angered many people.
Around the time that Mayhew’s investigation was released, in the fall of 2013, online discussion boards cropped up that attracted fire professionals and hobbyists. One blog still active today has tens of thousands of comments. Too many of them are overseasoned with vitriol or dedicated to conspiracy theories—somebody ordered the men to leave the ridge; a backfire sparked by a homeowner killed the crew; the hotshots were amateurs. These commenters often accuse Mayhew of being a conspirator in a government coverup. He calls the accusation patently false. But what bothers him is that some of those ideas have infected the fire culture, and he’s constantly having to correct dangerous misperceptions. “It’s comforting to think, ‘I never would have done that. I’m not like them,’” says Mayhew, who was a hotshot and still works as a firefighter. “They were just firefighters, and we’re just firefighters.”
Mayhew and I left the overlook and began hiking when the sun slipped below the Weaver Mountains and the peaks’ shadows stretched into the valley below. We followed the thin road that Granite Mountain took to their deaths. It was steep and rutted, and we both kicked rocks that tumbled downhill. We soon reached the point where the hotshots opted to drop off the ridge, through the canyon, and toward the ranch. We stood there for a moment. A turkey vulture rotated overhead. “Doesn’t it look like it’s right there?” Mayhew asked of the ranch we could see at the head of the canyon. “Like you could be there in five minutes?”
The uncertainty behind what drove those men, in view of that terrifying fire, to drop into a wickedly steep box canyon has generated the conspiracies that still haunt wildland firefighting today. In hindsight, it’s a hard decision to fathom. For his part, Mayhew tries to stay out of the swirling theories. He thinks the way to learn from Yarnell is to ask firefighters to put themselves in Granite Mountain’s boots and ask what could have lured them to make the same choice. On this point, he’s bullish. “They were trying to save lives,” Mayhew says. “They knew people were threatened down there. That must have weighed on them.”
Whatever it was that pulled them off that ridge, after years of making necessarily risky decisions on the fireline, Granite Mountain missed something on Yarnell Hill. And the numbers simply caught them. Mayhew grunted and set off down the hill, hiking toward 19 crosses five minutes from a ranch.