Fire

FIRE 1992

Late in the afternoon of July 26, 1989, a dry lightning storm swept through the mountains north of Boise, Idaho, and lit what seemed like the whole world on fire.

A dry lightning storm is a storm where the rain never reaches the ground. It evaporates in midair, trailing down from swollen cumulus clouds in long, graceful strands called virga. The electrical charges from a dry storm do not trail off before they hit the ground, however; they rip into the mountains like artillery. On July 26, 1989, lightning was hitting the upper ridges of the Boise National Forest at the rate of a hundred strikes an hour. Automatic lightning detectors at the Boise Interagency Fire Center were registering, all over the western states, rates up around two thousand an hour. By nightfall 120 fires had caught and held north of Boise, little one-acre blazes that eventually converged into a single unstoppable, unapproachable front known as the Lowman fire.

For the first three days Lowman was simply one among hundreds of fires that were cooking slowly through the parched Idaho forests. Around four o’clock in the afternoon of July 29, however, the flames reached some dead timber in a place called Steep Creek, just east of the town of Lowman, and the fire changed radically. The timber was from a blowdown two years earlier and was so dry that when the flames touched it, the entire drainage went up. The fire created its own convection winds, making the fire burn hotter and hotter until the fire behavior spiraled completely out of control. Temperatures at the heart of the blaze reached two thousand degrees. A column of smoke and ash rose eight miles up into the atmosphere. Trees were snapped in half by the force of the convection winds.

The fire rolled across Highway 21 and right through the eastern edge of town, detonating propane tanks and burning twenty-six buildings to the ground. A pumper crew was trapped at the Haven Lodge, and they hid behind their truck and finally stumbled out of the blaze an hour later, safe but nearly blind. The fire had attained a critical mass and was reinforcing itself with its own heat and flames, a feedback loop known as a fire storm. The only thing people can do, in the face of such power, is get out of the way and hope the weather changes.

Which they did, and which it did, but not until a month later, after forty-six thousand acres of heavy timber had been turned to ash.

I saw the site of the Lowman fire in 1992, three years afterward, when the ponderosa seedlings were already greening the hillsides. A roadside plaque said that eight million ponderosa and Douglas fir would be hand-planted by the mid-1990s. The plaque went on to describe how the land had been treated with enzymes so that water and microorganisms could penetrate soil that was now seared to the consistency of hard plastic. Thousands of flame-killed trees had been dropped laterally along the slopes to keep the land from washing away, and thirty thousand acres had been planted with grass and fast-growing bitterbrush. In a hundred years, more or less, the area would again look the way it once had.

I was driving a big, painfully beautiful loop from Ketchum, Idaho, around the Sawtooth Mountains and down the South Fork of the Payette River toward Boise. It was late afternoon when I drove through the Lowman burn, and the quiet darkness of the dead valleys depressed me. The West was well into one of the worst droughts of the century, and I was out there to see the wildfires that it was sure to produce. My idea was to go to Boise—where all the fire-fighting resources were coordinated—tell them I was a writer, and hope they let me on afire.

I pulled off down an old logging road and pitched my tent in a clear-cut. It seemed to get dark very quickly that night, and I cooked spaghetti on my camping stove and went to sleep listening to the weekend traffic die down on Highway 21. The Lowman fire, I’d heard, had burned so hot that Highway 21 had melted. There were places, I’d heard, where fire trucks had left their tread marks as they rushed from Boise to fight the flames.

In 1965 the U.S. government established the Boise Interagency Fire Center to coordinate the three federal agencies—the Bureau of Land Management (BLM), the Forest Service, and what was then known as the Weather Bureau—that were engaged in fighting wildfire in America. The Bureau of Indian Affairs, the National Park Service, and the Fish and Wildlife Service were added later, and the name was ultimately changed to the National Interagency Fire Center. Two years after BIFC was established, the Northern Rockies were hit with a catastrophically bad season that culminated in the Sundance fire in northern Idaho. BIFC managed to deploy thirteen thousand men and thousands of tons of supplies, prompting a study by the Office of Civil Defense, which was trying to figure out how to handle a similar crisis in the event of a nuclear war.

BIFC is located next to the Boise airport, across the interstate, south of town. The lobby is filled with the sort of display that, were you even vaguely inclined toward a job fighting fire, would make you move out west on the spot. There is a smoke jumper mannequin in full jump gear, including a wire face mask for when the jumper goes crashing into the treetops. There is a board with everything—food, medical supplies, tools—a jumper needs for forty-eight hours on a fire. There are color photos of air tankers dropping retardant and sheets of flame rising from stands of trees. One photo shows a fire in dense forest on the Umpqua National Forest in Oregon. “The total-timber jump spot,” the caption reads, “trees in this photo are between 80 and 125 feet tall. Five of six smokejumpers committed to this fire ‘hung up’ in trees, thus the rope carried in the leg pocket for a ‘let-down.’ The fire was stopped at a quarter-acre.”

A short, powerful man named Ken Franz—one of the Boise smoke jumpers, as it turned out—spotted me loitering in the lobby and came over to investigate. I told him I was interested in wildfire, and he motioned me into a cluttered conference room and sat me down at a long table to tell me the basics. Behind Franz was a map of the western United States that covered most of one wall. There were seven red circles on it around seven towns: McCall, Idaho; Boise, Idaho; Missoula, Montana; Redmond, Oregon; Redding, California; Silver City, New Mexico; and Winthrop, Washington. Franz turned and pointed to them.

“Those are the smoke jumper bases,” he said. “We are constantly getting sent from one place to another; you never know where you’re going to be the next day, or the next week. They just shift resources around to wherever the hazard is greatest.”

That shifting of resources, Franz explained, is what BIFC is designed for. Everything from government-issue paper sleeping bags to food to foam fire retardant to Ken Franz himself is shipped around the country, following fires, following thunderstorms, even following droughts. There are 410 smoke jumpers and perhaps 20,000 active and on-call fire fighters in the United States. Should a smoke jumper’s father die, say, or his house burn down, BIFC would know what state, what fire, what division, and what 20-person crew he was on. Should an air tanker go down en route from Denver to Missoula—one of hundreds of flights during a busy day fighting fires out west—BIFC would know what route it was taking and when it was supposed to arrive. The immense task of keeping track of all these things is accomplished at the logistics center on the top floor of the main BIFC building, across the parking lot. Across one wall of the room is a huge map of the country. Cardboard cutouts representing airplanes are moved around on it; cards representing fire crews are switched from “available” to “unavailable” slots. More detailed information is stored on a computer. In late August 1987 lightning started two thousand fires across the West that burned almost a million acres. In ten days 22,500 fire fighters and forty-five tons of supplies were deployed to fight the fires. BIFC accounted for every chain saw, every hard hat, every gallon of retardant.

“Smoke jumpers are considered an initial attack force,” Franz went on. “That’s a generic term for the first response to a fire. The classic situation would be a lightning-struck tree in a remote area where two guys jump in, fell it, buck it up—put out what amounts to a small campfire. Basically, the whole world’s a jump spot; within a mile of any fire you can usually find a very acceptable place to land in. On a big fire you have to start somewhere, so you jump a whole planeload and establish an anchor point, at the tail of the fire. You clear helispots for landing supplies, and you work your way around the sides of the fire.”

Smoke jumpers land with eighty pounds’ worth of gear, including two parachutes, puncture-proof Kevlar suits, freeze-dried food, fire shelters, and a few personal effects. Following them in cardboard boxes heaved out of the airplane with cargo chutes are chain saws, shovels, ax-hoe hybrids called Pulaskis, sleeping bags, plastic cubitainers of water, and dozens of other things needed on a fire. If there’s an injured jumper, a medical emergency pack comes out of the plane. If it’s a fast-moving fire, the crew can call for boxes of explosives that can blast an instant fire line in the forest duff. The list of what can be thrown at a fire is endless—and expensive. A more cynical view, popular among many, is that the government puts fires out by throwing money on them until it starts to rain.

Not much of the money, however, goes to the fire crews. A jumper makes about $8.50 an hour. If the fire is uncontrolled, as, since smoke jumpers are initial attack, it almost always is, the crews get another 25 percent hazard pay. If they work overtime—again, almost a sure bet—they get time and a half. The jump itself has been ruled as simply another way of getting to the fire, like a bus or a pickup truck, so jumpers get straight pay when they leave the airplane and time and a quarter when they hit the ground. If they are injured on the jump, however, and don’t make it to the fire, the hazard pay does not kick in. From fifteen hundred feet it takes about a minute and a half to reach the ground with a parachute. At $8.50 an hour, that’s about 21 cents.

“In a good year you can make almost thirty thousand dollars,” said Franz. (As with all fire fighters, a “good” year is a year with a lot of fires; a “good” fire is a fire that isn’t brought under control too quickly.) “Under twenty thousand is more typical. That’s for six months. The rest of the year we sew.”

They sew everything: harnesses, fire line packs, jump bags, even little duffels with the BIFC flame logo on the side. They do it to save the government money, they do it because they’re better sewers than most manufacturers, and they do it to keep themselves employed. The only thing they don’t sew are the parachutes. Some jumpers are certified to make repairs, but the chutes themselves are bought from a manufacturer. The parachutes the BLM uses cost a thousand dollars apiece and are expected, with upkeep, to last at least ten years. They were of a design invented by a French kite maker in the early 1900s. They are called Quantum Q5 Ram Air parachutes.

“Ram air means there are cells that fill with air,” said Franz. “They make the canopy so rigid you could walk across it. You could also put a line on it and fly it like a kite; in Alaska, jumpers fly their chutes like kites. You steer with toggles and have a forward speed of twenty miles an hour. It’s a very high tech delivery system for a very low tech job; once we hit the ground we’re just fire fighters. Afterward we have to pack ten miles or more, to the nearest helispot. Our gear weighs over one hundred pounds, and usually we’re not even on trails; it’s harder work than fighting fires. It keeps you honest.”

Honest means capable of enduring a training regimen that used to weed out 30 percent of the preselected men at the training camp (overwhelmingly men, but not entirely). Rookies are considered the fittest and most perfectly trained because they have endured boot camp most recently: three hours of workouts a day, a jump simulator called the Mutilator, an array of courses and tests that virtually guarantee you’ll pull your ripcord after jumping out of the plane. Overwhelmingly, it works, though not always. In 1991 a jumper in Montana was killed because he didn’t reach for his ripcord until “ground rush,” when it was too late. The entire thing was caught on video because it was a training jump. The consensus was he froze.

“The biggest hazard is probably the fire itself,” Franz told me. “Felling burning snags, logs rolling down hillsides. Jumping is usually a relief. It’s hot in the airplane, and sometimes you feel sick; then suddenly you’re totally focused on what you’re doing. It’s a little dreamlike.”

After our talk Franz took me for a quick tour of the jump loft. He showed me the rigging room where the chutes are packed, and the sewing room, and the weight room. Afterward we returned to the conference table, and he popped a short tape into the VCR. It was quick and unprofessional but highly dramatic. It showed a jump crew working a fire at night, right on the line. At one point a sawyer was cutting down a huge ponderosa, and his saw was halfway through the trunk when flames started pouring out like liquid. The tree was hollowed out by fire, it turned out, and was drawing like a chimney. The sawyer kept cutting; the flames kept spurting; eventually the tree fell.

As I left, I asked Franz—against all hope—if there were any fires around for me to see. He told me I’d just missed a good one. An older couple from Pennsylvania had been towing a car behind their RV, and the car got a flat tire; sparks started a fire front two miles wide. Six thousand acres, a million dollars to put out.

“I suspect the government will try to collect too,” he said. “Try the dispatch office; they’ll know what’s going on.”

The dispatch office for the Boise National Forest was a trailer east of the airfield. “I just came from BIFC,” I told the young woman behind the desk. “I’m a reporter. Are there any fires?”

I felt a little bad asking the question. She didn’t blink. “Seven hundred acres as of midnight last night,” she said, spreading a map of Idaho out on the table. “Lightning-started, wind-driven, with three helicopters and twenty-two crews. It’s called the Flicker Creek fire. They’ve called in a type one overhead team.”

A type one overhead team is called in only when a fire is really bad or is expected to get really bad. The Flicker Creek fire was in steep terrain with extremely dry fuels and strong winds. Steep slopes help a fire because uphill fuels get preheated; winds help a fire because they make it burn hotter and push it across the land. A seven-hundred-acre fire could jump to seven thousand or even seventy thousand in no time at all.

An hour later I was driving north on Highway 21 in my green fire-retardant Nomex pants and yellow fire shirt. In the back seat were a yellow plastic hard hat and a fiberglass and aluminum fire shelter. The shelter is a pup tent that comes in a small pouch with belt loops. It reflects radiant heat, reducing what would be a 1,000-degree fire to 120 degrees or so. I would be assigned a public relations person when I got to the fire camp, the ranger told me. I would be fed and I would be lodged in a tent if I didn’t have one. Tomorrow morning a helicopter would take me into the fire line.

I breezed past some bored Forest Service guards and turned off Highway 21, into the hills.

Flicker Creek is one of hundreds of small creeks that cut through the steep, dry hills of the Boise National Forest. Most of the land is grass and rocks and sagebrush, with heavy stands of ponderosa on the north slopes and in the drainages. Flicker Creek empties into the North Fork of the Boise River, which quickly joins the Middle Fork and continues on to fill the Arrowrock and Lucky Peak reservoirs. The entire West was seven years into one of the worst droughts since the 1870s, so both reservoirs were severely depleted. Arrowrock had been reduced to a muddy brook that you could practically jump across.

After twenty miles of rough driving, the road leveled off along the North Fork of the Boise. There was plenty of water up here—or so it looked—and the river was fast and lined with big, open stands of ponderosa. The fire camp was in a huge meadow called Barber Flats that ran alongside the North Fork of the Boise. Hundreds of bright nylon tents were pitched in the yellowed grass. A helicopter thumped over a ridgeline, trailing a retardant bucket. Water trucks rumbled back and forth, spraying the dust down. Hotshot crews came and went, Indian file, or slept in the shade, or sharpened their tools. Some were black with dirt; others looked as if they’d just arrived. They all had on the same green and yellow Nomex that I wore and big lug-soled boots.

I parked my car between the trucks and water tankers and searched out the information desk. The public affairs people knew I was coming, and I was pointed toward a large, deep-voiced man named Frank Carroll. “You’ll need boots if you want to go out on the line,” he told me. “You’ll need water bottles; you’ll need food; you’ll need gloves. I’ll set you up after dinner. You can pitch your tent anywhere you like. People get going around five in the morning; make sure you’re at breakfast and ready to go by then.”

I thanked him and went off to get my gear set up. All around me, big, lean men and a few women went about their duties. I pitched my tent in tall grass behind a cabin that served as a command post and then wandered over toward the catering tent. Behind it was a full-size truck outfitted as a kitchen. Hotshot crews passed by it in line, taking plates of food from the young woman behind the window and then sitting down at folding tables under a canvas army tent. The woman was pretty and had a sheath knife on her belt. I tried to pretend I belonged there, and she loaded my plate up with steak and carrots and mashed potatoes and salad and two slices of white bread.

I took a seat by myself at a corner table and watched the crews come and go, talking loudly, eating fast. Most of the fire fighters were young white men, sinewy and unshaved. There was a scattering of women among them, but the women were treated—as far as I could tell—no differently from anyone else. I didn’t think they would be discriminated against so much as subjected to one form of gallantry or another, but they weren’t. Everyone seemed to be too tired and hungry to notice the opposite sex. (This turned out to be emphatically untrue.) Furthermore, beneath the baggy clothes and grimy faces it was hard to tell who was what.

The Indians and Latinos generally had their own crews. It was a reflection of demographics more than anything else: Twenty men from Browning, Montana, are likely to be Blackfeet, not white; twenty men taken off the farm crews in the Snake River valley are likely to be Latino, not Indian. It’s the Indian crews, the caterer told me, that can really clean out a food truck. “They’ll eat anything that’s not nailed down,” she said. The convicts eat sweets and spicy foods because they can’t get much of that in prison. The white ’shot crews are the most health-conscious: They eat a lot of fruits and vegetables; some won’t even touch meat.

That was hard to imagine, because the government food contracts were defined by how much protein—meat, in other words—was provided per person per meal. Everyone gets four ounces of meat at breakfast, seven ounces at lunch, and ten to sixteen ounces at dinner. Everything else—vegetables, grains, fruit—was considered a condiment and didn’t figure into the equation. It’s a lot better than it used to be, though. Back in the dark ages of fire fighting—before women, before showers, before Nomex—the crews subsisted mostly on ham.

Ham and eggs, ham sandwiches, fried ham. Catering trucks were essentially big meat lockers with ham hanging in them, and maybe some Wonder bread. Back in those days the hotshots wore T-shirts that said, “When forests burn, pigs die.”

After dinner I sat in on the planning session. It was held under a ponderosa pine by the dirt parking area. The entire overhead team was there, identifiable by the fact that they weren’t dirty and weren’t wearing Nomex. They were trained together and used on the Flicker Creek fire as interchangeable parts of a network called the Incident Command System. The system is based on the idea that any person trained for a certain job—logistics chief, information officer, helibase manager—can perform that job for any agency, in any situation. Overhead teams are made up of people from a dozen different government agencies and are pulled in from all over the country. You might find an incident commander from Georgia and an air operations branch director from Colorado and a safety officer from the next town over. There are seventeen type one overhead teams in the country, and they mainly fight fires, but they have also been effective on other catastrophes: oil spills in Alaska; hurricanes in Florida; earthquakes in Mexico. An overhead team was sent to clean up the Valdez oil spill, for example, and the system worked so smoothly that it was copied by both the Exxon Corporation and the U.S. military. A fire camp with an overhead team, in fact, can put two people in the field for every one person acting as support—a ratio roughly twenty times as efficient as the military’s.

I went to bed when the five hundred fire fighters did, at dark. The only noise was the continuous rumble of the generators. The planning session had brought bad news, in a way: The fire was cooperating almost too quickly. A thirty-acre spot fire had started in light fuels on the south front but had been contained by three crews. Seven type two crews—less experienced than hotshots and usually used for mopping up—had cut line all the way down to the river, farther than expected. The winds were dying down, and unless they picked up again, the fire would be contained within days.

A fire camp is never completely still. All night long I was aware of the movement of men. They walked past, packed equipment, coughed, spat. Around four in the morning the sounds were so continuous that I woke up even before my watch alarm went off. It was still dark, and the camp undulated with human forms and occasional headlamps. Crews were packing their line gear, drifting toward the catering tent, clustering around the big stand-up kerosene heaters set up at intervals in the field. It was cold, maybe in the twenties. I crawled out of my sleeping bag and pulled on several sweaters and my boots—I had to wear leather on my feet in the helicopter, for some reason—and hustled toward the lights of the tent.

The person assigned to me was Bill Casey, a type two safety officer from the Boise area. He was a strong, clear-eyed man in his late forties who directed the local Bureau of Land Management district and was also qualified to command a type two overhead team. (A type two team handles smaller fires, but operates the same way.) He had bagged thirty elk in the past thirty-two years of hunting them, he said. His father hunts with him and can still shoulder fifty pounds of elk meat, at age seventy-one. Casey is part Blackfeet; he has dead-straight gray hair and brown eyes and a handsome, open face.

“We’re a little overstaffed because the fire didn’t do what we expected,” he admitted as we sat in his truck at the helibase.

“What did you expect the fire to do?” I asked.

“Well.” He chose his words carefully. “The guys don’t want to see the forests burn, but on the other hand, they want to have a good, productive summer. They like to have two to three weeks on a fire and then move on to another one. In that sense, it’s disappointing to have the fire lay down so fast.”

We were waiting for our flight into the fire line, and the heater was going full blast in the truck. In the clearing, men were checking the helicopters and writing up flight manifests. The first few days of a fire are usually disorganized to the point of chaos, Casey said. Then it gets better. Casey took advantage of the wait to tell me more about how fires are fought—not from the field but from the office.

“BIFC is just a logistical center that responds to needs in the field,” he said. “Suppose Unit X has a fire. As long as it doesn’t escape the initial attack, BIFC is not involved. If it does escape, then a regional coordination center gets involved. If the regional center doesn’t have enough resources to fight the fire, then BIFC steps in. We’ve had a lot of fires already this summer, and this is the first one where BIFC has been involved.”

Initial attack, he said, comes in many forms. Smoke jumpers are initial attack. Helirappelers are initial attack. Hotshots can be both initial attack and extended attack. Air attack—retardant drops from planes and helicopters—can also be initial or extended. The idea of initial attack is to hit the fire hard and early so that you avoid the expense of an extended campaign. If the attack crews can’t contain it, then an overhead team is assembled and put on the fire. A really big fire will suck in ’shot crews from all over the country. There are a total of sixty nationwide. If that’s not enough, or if they’re needed elsewhere, then type two crews that are trained in fire but often do other work for the Forest Service and BLM are deployed. Farther down in the barrel are convict crews, laborers from the Snake River valley who collect at the center of town when sirens go off, and pickup crews. Pickup crews usually just consist of people with sturdy shoes and in good enough health to pass the training course. When there are pickup crews on a fire, it’s a really bad fire. There were no pickup crews on the Flicker Creek fire.

After about an hour the helicopter was ready. It was a Bell Jet Ranger, rented from a private contractor in Arizona for two thousand dollars an hour. The huge Croman logging ships cost three times that and can sling twenty thousand pounds of retardant in a bucket beneath them. We were read the safety procedures by a helitack crewman and then asked how much we each weighed, with gear. The answers were entered into the flight manifest to calculate how much the helicopter could carry once we were out of the ground effect. Once a helicopter rises about a certain height, the downwash from its rotors no longer has the ground to push against, so it can carry less. If you combine the altitude, the air temperature, and the relative humidity, you have a figure called density altitude that determines how much a helicopter can carry. The retardant buckets, called Bambi buckets or Sims buckets, depending on their manufacturer, could even be cinched down to change the size of the container to compensate for variations in density altitude from day to day.

The crewman went over the instructions again: Keep your head down when you come and go from the helicopter. Never leave a helicopter uphill. Never go toward the rear of the helicopter. Never go anywhere that the pilot can’t see you. If you violate any of these rules, the attendant is authorized to use force to get you to comply.

“I’ll tackle you,” he said.

It was the first flight of the day for this helicopter, so the rotors weren’t turning when we boarded. Gloves on, helmets and goggles on, sleeves rolled down. The attendant made sure we had buckled our seat belts, and then he closed the doors and the rotors heaved into motion.

“By the way, why couldn’t I wear sneakers in the helicopter?” I shouted to Casey.

“Because if we crash, the nylon melts to your skin,” he yelled back.

In front of me, the copilot had a flight helmet that said, “Crashing Sucks.” The technical term for crashing is hard landing. One pilot said that half the pilots he knew had done it. Some had been killed; most hadn’t. Many helicopters these days are designed to autorotate, meaning that they don’t whistle straight downward if the engine fails. They do an awkward imitation of a glide. Most of the Boise National Forest is rocky and steeply angled, though, so it was hard to imagine its turning out well one way or the other.

The Bell Ranger lifted off, and within minutes we had put the camp behind us and I could see smoke coming off a distant ridge. It was not the furious orange blaze I had imagined; it was a 750-acre smudge fire that, if touched by wind or left unattended, could spring to life and go ripping off into the heavy timber north of here. Helicopters were shuttling back and forth from Barber Flats, where they dipped their buckets in a retardant tank. The retardant was rust-colored, and you could see it splotched on the hillsides. Below us, one of the big Cromans was easing down into a steep valley to fill its Bambi bucket in a stream. (Hotshots are fond of saying that fish get scooped up in the bucket and can be found flopping on the hillsides. No one had actually seen this happen, but the story became so widespread that the Fish and Wildlife Service hired a helicopter to see if its pilot could catch fish on purpose. They failed.) The Croman, Casey yelled, carries a bucket that is too big to fit in the retardant tank back at Barber Flats, so it has to use water. It will go back and forth all day between the stream and the fire, directed by ’shot crews with radios. The pilots are so accurate that they can hit individual trees.

We pounded in low along a ridge and then rose and banked and circled and came in again. I could see yellow-shirted hotshots below us, shrouded in smoke. The fire was down in the valley. Several crews were downslope, strung out along the creeping line of black, and several more crews had just landed at the helispot on the ridge and were waiting for orders to proceed. The helicopter made one more approach and then settled down uneasily onto a ridge that plunged steeply away on both sides. The ’shot crews crouched down against the rotor wash, and the helispot attendant, dressed in a green flight suit and hard hat and goggles, flattened himself behind a boulder as if someone were shooting at him. When the aircraft had set itself down but with the rotors still thumping, the attendant scurried forward in a crouch and opened my door and took me out by the arm. Then he returned to get Casey. Occasionally people back up into the tail rotor or walk uphill into the main one, so there is always someone at the helipad to act as an escort. I dragged my pack over behind some bushes and knelt with my head turned away. The helicopter clattered up over our heads and then continued on to shuttle more overhead around the fire.

Already the sun was hot, and it was going to be a long day of mop-up. Hotshots consider themselves elite, and a day spent cold-trailing—going through a burned area and putting out every ember—is not their idea of fun. It probably isn’t anybody’s idea of fun, but the type two crews get stuck doing it because by then the ’shot crews have usually been flown to some new emergency.

Today, however, the fire was not cooperating, as Casey said. Unless the wind picked up, everyone would be cold-trailing. The hotshots untaped the axes and shovels and Pulaskis that had been protected for the helicopter flight, and fueled up the chain saws—big Stihl 44s and 56s that the sawyers carried over their shoulders with a pad strapped to their suspenders—and took last swigs of water. It was hot on the ridgetop, but it was going to be ovenlike at the bottom of the valley, where the fire was burning and the wind couldn’t reach. Crew by crew, they reluctantly filed off downhill: the Helena Hotshots in their bright pink hard hats; the Flathead Hotshots; the Chief Mountain Hotshots from northern Montana, Blackfeet to a man. Casey motioned to me, and we started off down the slope, following a fire line that the type two crews had put in the day before.

A fire line doesn’t look like much: just a strip of ground scraped down to mineral soil that snakes around the perimeter of a wildfire. There are different kinds of fire line; the one we were on was called direct line, or hot line. It was several feet wide and cut right along the edge of the fire—“one foot in the black,” as they say. When the fire front is really flaming, the crews cut indirect line anywhere from a few feet to several hundred yards away. Often the area in between is then burned out with torches, called fusees. A burnout uses up the fuels between the line and the fire front, effectively creating a fire line up to several hundred yards wide. A burnout is different from a backfire, which is set deliberately to eliminate huge swaths of fuel in the path of an advancing fire. Often that is the only way to stop a crown fire that is throwing embers miles ahead of it; often, of course, a backfire becomes a disaster of its own. A burnout can be ordered by a crew boss; a backfire can be decided on only by the incident command team.

All handline—as distinguished from catline, which is cut by bulldozers—is made the same way. First, a line scout goes ahead of the crew and flags the route with red tape, taking advantage of anything—streams, rock outcrops, ridgelines—that won’t burn. The sawyers follow after the line scout, cutting everything from knee-high sagebrush to 150-foot trees. Each sawyer is assigned a swamper, who pulls the brush out of the way and throws it well outside the line. After the swampers come the rest of the crew, who use rakes, shovels, Pulaskis, and even their hands to scrape the duff clear and remove every shrub or root that crosses the fire line. Untouched forest or range enters at the beginning of this line, and a fuelless strip emerges out the other end. It is supposed to stop fire, and despite its appearance—narrow and insignificant in the midst of such huge geography—it generally does.

Fire line is built and measured in sixty-six-foot sections called chains, a unit of measurement that dates back to the early days of surveying. There are 80 chains to a mile, and hotshot crews should be able to cut 20 chains—a quarter of a mile—of fire line an hour. If it’s an emergency, the crew should be able to continue at that pace all day, all night by headlamp, and even all the next day. The unofficial record is sixty-seven hours, set by a California crew boss who had also gone thirty days without a shower. Technically both are in violation of agency policy.

For the time being the fire line served as a very good trail on the steep hillside, and I followed in Casey’s dust down toward the crews below us. Fifteen hundred feet farther down was the river. Casey’s job, as safety officer, was to walk around all day watching people work and talking to them about potential problems. Do they have a safety zone to retreat to if the fire blows up? Are people being posted as lookouts on each crew? Is everybody wearing a shirt? Most fires are slow-moving, and fighting them is closer to hoeing a garden than being in combat, but when fires blow up, they move with awesome fury, and if people aren’t prepared, they die.

“Seventy-two people were overrun at the Butte fire in ’85,” said Fred Fuller, part of the overhead team of one sort or another (“We’ve got layers upon layers of us out here,” he admitted). He was a lanky, soft-spoken man who was accompanying Casey and me on our rounds for a while. “They were in their fire shelters for an hour and a half. They lived, though—even the Cat operators who didn’t have fire shelters. They had to crawl under their machines. But they lived.”

Cat operators suffer some of the highest fatality rates on fires because they are reluctant to leave their machines until it’s too late; on the Butte fire they had to be dragged off their machines by other fire fighters. Burnovers are considered catastrophes even if no one gets killed, and the people who survive are given counseling within twenty-four hours. By all accounts they are terrifying experiences; that close, a fire storm is as loud as a jet plane taking off, and many of the people trapped in their shelters don’t even have radio communication. All they can do is wait and try not to let the convection winds tear the shelters off their backs.

Bob Root, a young severity crew foreman working downslope from us, put it bluntly: “If someone has to get into a fire shelter, then someone else fucked up. I mean, really.” Root has fought fire for seven years, since he was eighteen. He was a student at the Colorado State School of Forestry and also a member of the local sheriff’s search and rescue team. His severity crew, hired in advance because of severe fire conditions, was fire-ready around the clock and could be on their way within an hour. Root had straight straw-blond hair and a sunburn and dark Bollé glacier glasses. Realizing that I was as good an excuse for eating lunch as any, he sat down and pulled a brown paper bag out of his line pack.

“Probably the worst thing that’s happening now is what’s called urban interface—that is, houses in the forest,” he said. “When it’s a question of saving structures rather than just trees, we’re more likely to take risks. Basically, if you’re protecting a structure and the fire’s coming at you, you don’t retreat. You stay put and try to work the fire around either side.”

Root ate methodically, clearly in no hurry to resume cold-trailing. His bag lunch, delivered by helicopter, seemed geared toward the taste of grade school boys: bologna sandwiches, candy, cookies, more candy, and a couple of apples. As we talked, one of the Cromans came in low over us and started to ease itself down over the helispot. Dangling below it was a car-size bladder bag of water called a blivet. The pilot, looking down out of his side window, placed the blivet gingerly on the crest of the ridge and then released it from the tether. Root didn’t take his eyes off it.

“Sometimes when they punch off, the load goes downhill,” he explained. I realized that we were, in fact, directly below one and a half tons of water. He took a last look at it as he ate his sandwich.

“This is a pretty unexciting day, but last night was interesting. We were at the bottom of this slope—you never want to be above a fire—and rocks and trees were burning loose and just rolling past us. You could hear them coming. Someone would yell, ‘Rock!’ and then there’d be total silence while people tried to spot it and get out of the way. Logs would roll to the bottom and ignite the hillside, and the fire would come right back up at us. The night went by fast.”

According to Root, all over the West, fire conditions were about as dangerous as they could get. The drier the fuels, the hotter they burn and the faster the fires spread, and fuel moisture levels that should be at 15 or 20 percent are down in the single numbers. Low relative humidity and unstable air (wind) compound the problem. Fires are generally slow-moving creatures, moving a few chains an hour. But sometimes they can explode up a hillside or across a canyon, and the mountains all around us, scorched by seven years of drought, were as likely as they’d ever been to produce such behavior.

“I’ve never seen it so dry,” said Root, fingering the yellow cheatgrass around us. “One little ember in this stuff, and it ignites; last night every ember was taking. The fuels in Washington and Oregon are drier than what you’d buy at a lumber store. Now all it takes is one lightning strike in thousand-hour fuels for it to catch. That’s almost unheard of.”

Thousand-hour fuel is a piece of wood between three and eight inches thick. The thousand hours mean that if the fuel were completely saturated with water, it would take a thousand hours for it to lose 63 percent of its weight through evaporation. Conversely, if it were bone-dry, it would take a thousand hours to soak up 63 percent of its weight in moisture. Sixty-three percent is used as a benchmark because it is midway between two points—above 78 percent and below 58 percent—where moisture absorption or evaporation happens in a predictable, linear fashion. In that middle range, however, wood gains or loses moisture in a very complex way, and 63 percent is at the mathematical center point of that nonlinear range. Grass and twigs dry out or saturate almost immediately—one-hour fuels. Sagebrush and other small growth are considered ten-or hundred-hour fuels. Thousand and ten-thousand-hour fuels include everything else, up to ponderosa pine with six-foot diameters. It is very rare for thousand-hour fuels to be as dry as one-hour fuels, but they are. Everyone was worried. As one hotshot said, “There are no small fires anymore.” Everything that ignites tends to explode.

Fuel moisture levels can be established by field tests or by extrapolations based on information from the National Weather Service. The fuel moisture level is factored into an index that includes weather conditions, wind speeds, fuel loads (the total oven-dry weight, per acre, of all the fuels in the area), and drought conditions. The information is processed by the National Fire Rating System, which determines the fire risk for every climactic region in the country. It predicts such things as fire intensity, the likelihood of lightning strikes, and what is called the ignition component, the probability that a single firebrand landing in dry fuels will start a fire that requires a response. In addition, precise and highly accurate “spot” forecasts can be derived by the National Weather Service for areas as small as a quarter acre. A spot forecast may state that the temperature range for a specific canyon in the Boise National Forest will be between seventy-eight and eighty degrees; the humidity between 12 and 14 percent; and the winds ten miles an hour. Such information is crucial when ’shot crews are deployed in high-risk situations.

At the macro end of the spectrum is the general state of drought. In 1988 the whole country suffered through a brutally hot summer: Barges were running aground on the Mississippi River; railroad tracks were warping in New Jersey. The drought culminated in the fires at Yellowstone National Park. From late June until early November fires burned a total of 1.2 million acres across the West, more than half of which was in Yellowstone itself. On August 20 alone—Black Saturday—seventy-mile-per-hour winds fanned three-hundred-foot flames through a total of 165,000 acres. Although much of the country returned to normal that winter, the drought in the western states never really ended. Drought-stressed trees died by the thousands. Dry lightning storms continued to ignite huge fires. Fuel moisture contents dipped to 2 and 3 percent. The Palmer Drought Index—a water accounting system that measures water gained and lost in a given area—indicated that the entire West was entering the driest period since records had been kept in the 1870s. On the Palmer Drought Index, 0 is normal, +4 is extreme flooding, and-4 is extreme drought. During the summer of ’88 Yellowstone registered -5.8. During the Flicker Creek fire the National Weather Service declared the Boise area to have hit -6.5.

Out here that is not an abstract figure; it has the ability to stun. I repeated it to a division superintendent on another fire, and he jerked his head back as if he’d been slapped.

Root finished his last candy bar, and we both stood up. I thanked him for the talk. Helicopters thumped continuously overhead with the Bambi buckets, and ’shot crews raised plumes of dust on the steep hillsides. Casey angled slowly uphill toward me in the high August sun, and after he arrived, we contoured across the ridge and headed for the west side, where the Chief Mountain and Flathead ’shot crews were cutting snags and putting out spots above the river. Snags (standing dead trees) are routinely cut down whether they pose a fire threat or not; some are completely hollow inside and standing on a virtual toothpick. When they let go, it is without a sound, and occasionally there is someone standing underneath. At the Red Bench fire in 1988, a hotshot was killed by a snag while waiting for a bus by the side of the road. “Death comes from above,” one hotshot warned, explaining why he walked around with his head tilted up.

We stopped to talk for a while to the Flathead crew, two of whom had been at the infamous Dude fire in Arizona when a convict crew were burned over in their shelters. Melissa Wagner, now reclined in the cheatgrass eating lunch, recalled hearing the screams of the convicts over the radio as they died. One of the survivors left his fire shelter too early and emerged from the flames with his hair smoking and 47 percent of his body burned. Six others died. Wagner kept fire fighting because she needed a way to pay for law school.

Casey and I continued toward the Chief Mountain crew, visible on a distant ridgeline with smoke coming up behind them and helicopters circling above. Halfway there, trudging through the dust and baking heat, Casey suddenly jumped back and almost knocked me down. A big rattler was coiled on a rock in the middle of the fire line. It didn’t move, and it didn’t rattle; in fact, it was headless. We reached the Blackfeet crew twenty minutes later. They were sitting in the shade of a small ponderosa, eating their lunches; way downhill, a solitary tree was torching.

“So, you guys know anything about that rattler back there?” Casey asked.

There was silence. Every man in the crew looked off in a different direction.

“Rattler?” Glen Stillsmoking, the crew boss, finally said.

“Sons of bitches,” Casey muttered, shaking his head but with a smile.

We sat down in the shade to drink some water and take in the view. One of the crew pulled an obsidian arrowhead from his pocket; he said he’d found it at the helispot at Barber Flats. Stillsmoking leaned in to take a look. “We were a pretty hostile people,” he admitted, looking around at his crew. “We ran Chief Joseph off; we were the last to settle down. But now we make good fire fighters. My father was a fire boss; I wanted to go to flight school, but fire just drew me away. We’ve been to Alaska, Florida, paid vacations—that’s what we call work—everywhere. At the safety briefings in Florida they told us to watch out for the alligators.”

Downslope, a pair of sawyers—sawdogs, they’re called—were dropping snags; the sound of their Stihls reached us as a puny whine. A lone ponderosa burned lazily behind us a mile away, and the river glinted far below us in the canyon. This was probably as peaceful as it got on a fire line, and the crew didn’t seem in a particular hurry to ruin the moment. I leaned back and tried to stop wishing the fire would do something big.

The season blew up a few weeks later. I was back on the East Coast when I got a call from Frank Carroll, the information officer for the Boise National Forest, who said that lightning strikes were now starting fires by the dozen at the upper elevations. A twenty-year-old fire fighter had been killed by a snag at the tiny Cascade fire; it hit her so hard that her hard hat was driven into her head. A smoke jumper broke his pelvis while landing on the Red Mountain fire. The town of Cuprum had to be cleared of all thirty residents when the Windy Ridge fire detonated into a five-thousand-acre blaze in one afternoon. A Diamond Mountain hotshot on the Horsefly fire was knocked five hundred feet down a hillside by a flaming log because he had shoved a friend out of the way before trying to dodge it himself. A total of eleven thousand people were on the fire lines at one time, Carroll said, and water levels were so low that Boise-area farmers had been shut off two months early.

A big runaway fire was almost inevitable, and it finally hit at the end of August. On the afternoon of Wednesday, August 19, a thunderstorm swept past Boise and lightning ignited the rangeland east of town. The fire quickly overwhelmed the BLM engine crews sent to deal with it and made its way up into the Boise foothills, leaping the quarter-mile-wide canyon of the South Fork of the Boise River with ease. By then it was in steep terrain and flashing through the grass and sagebrush almost faster than a person could run. Temperatures were in the nineties, and the relative humidity had bottomed out at 5 percent. Conditions were so fast that the fire encircled the town of Prairie and almost torched it, forcing the evacuation of all one hundred residents. The only store in town—complete with a hitching post and a bar—was selling T-shirts that said: “We interrupt this marriage to bring you the fire season.”

It was called the Foothills fire, and I got on it ten days after it first ignited. Snags in the timber were dropping at an estimated rate of forty an hour, prompting overhead to pull crews off the lines at night. This in turn prompted locals to say that the fire fighters weren’t working hard enough. I was assigned to the Union Hotshots out of La Grande, Oregon. Not only was Union one-third women, but its crew boss, Kelly Esterbrook, was one of only ten women ever to have made it through the brutal smoke jumper training course. The Union Hotshots were one of three crews guarding a strategic section of the fire line and if temperatures stayed high and the wind picked up, they would be right in the middle of things. The fire could roll right across the handlines and into a big stand of diseased ponderosa in the high country, and if it did that, it would be virtually unstoppable.

A public information officer named Karen Miranda had been assigned to take me up the fire line to meet the Union Hotshots. I picked up my Nomex clothes and fire shelter at the Forest Service headquarters in Boise, bought notebooks and blank cassette tapes, and raced east toward the town of Prairie. It was thirty miles away across a huge swath of dead black rangeland. I had to make the fire camp before three o’clock, when the helicopters started making their evening runs out to the crews, and I got in with a half hour to spare. When I walked up to Ed Nesselroad, the senior fire information officer, he was telling someone about a live elk that had been found with its eyes burned out. He quickly rounded up Karen Miranda, and by late afternoon we were strapped into the canvas seats of an Evergreen helicopter and waiting to be shipped to helispot six—H-6, as it is called—northeast of Prairie.

We got the same warnings as before, except that this time we were shown the emergency shutoffs in case we crashed and the pilot was unconscious. (Never get out of a downed helicopter with the rotors still turning.) We lifted off and immediately could see a cauldron of smoke boiling up out of a valley. Dozens of helicopters flew in and out dumping retardant, and away to the south stretched an endless carpet of blackened rangeland. Trees torched below us, and the sweet, sharp smell of smoke filled the cabin. The helicopter made several low passes at H-6 and finally settled down after the helispot crew had cleared some cargo out of the way.

H-6 was on a ridge just below a vegetation line where ponderosa gave way to alpine fir. Just over a hill was tiny Smith Creek Lake, cupped in the topography like a jewel in some huge brown palm. Tents and bedrolls were scattered through the ponderosa. Hotshots lounged with books or talked in small groups or just sat and stared. A slingload of gear waited to be sorted below the helispot, and near it was a huge pile of cardboard boxes and plastic buckets. It was our food for two days.

An inflamed red sun was setting over the ridge, and smoke was pumping out of a valley to the west of us. Occasionally a tree torched on the ridge and then died down, lighting the campsite—it was dark by the time we ate—with a dull glow. Hotshots stopped eating and turned to watch. These were people who couldn’t even remember how many fires they’d been on, I thought. Still, they couldn’t keep from looking at open flame. There are stories of crews getting overrun by fire because they were too mesmerized by it to run away.

The camp we were at was termed a spike camp, and the hotshots here—sixty of them plus several helitack crew members—were said to be spiked out. That meant that they were established in a roadless area and supplied by helicopter with food, tools, and paper sleeping bags (marginally warm but washable and reusable). According to Forest Service policy, hotshots should not be spiked out for more than two days in a row. One level less comfortable than a spike camp is the coyote camp, and hotshots are not universal in their love of coyote camps. Coyoteing, as it is called, means dropping in exhaustion wherever you happen to be when it gets dark. Because hotshots have only their line packs when they fight fire, they are usually caught without food, sleeping bags, extra clothes. If it’s cold, they will make a fire in the black—the burned area—and huddle around it all night. If it’s really cold, they might decide to keep building line simply to keep warm. For food, they might have thought to pack some military MREs (meals ready to eat). If not, they go hungry.

The trees torched intermittently all night, I would wake up and see their glow. The sound of flames consuming pine trees one by one easily carried the mile to camp. I tried, and failed, to imagine what it would be like to be burned over. In 1910 a fire storm overtook a group of forty-five fire fighters who were trying to escape an Idaho fire called the Big Blowup. The leader, Edward Pulaski—who later manufactured the tool that bears his name—led them into an abandoned mine shaft and had to keep them there at gunpoint because they were so terrified. Five men died; the rest emerged several hours later, burned and dazed. They had survived a fire storm of such ferocity that entire hillsides of timber had been flattened by fire and convection wind.

The hotshots were up well before dawn, as usual, rustling about quietly with their head lamps on. It was very, very cold at that altitude at that hour, and I put on everything I had and wandered down the hill to the kitchen area. There would be a briefing at six while people ate, and then they would attack the fire. The smell of smoke still permeated the camp, but the fire had quieted down during night, and trees no longer torched along the ridge. Fires generally “lay down” after dark because the temperature drops, meaning that the relative humidity rises. People poured themselves cereal and coffee and sat on the hillside, eating and listening to the morning briefing.

The division supervisor, a gruff, stocky man named Fred Bird, stood on a crate in the half-light and projected the plans for the day out across the mountainside. “Okay, we’re gonna try to hold our own on that ridge,” he said. “We’ve got good air support, and they’re going to work us here all day today and then bump us back to camp. Tomorrow they’ll get in some type two crews to hold the ridge and send ’shot crews up to the North Zone to chase spots down that canyon.”

One hundred and seventeen miles of fire line had been built, there were only ten more miles to build, and overhead was saying that the fire was nearly contained. Everyone knew that one good spot into some timber could start the fire running again, though. Bird’s briefing was short and to the point, and sunlight was just touching the upper ridges when the three crews shouldered their packs and started off up the hillside. I fell in with the Union Hotshots, followed by Karen Miranda, and we worked our way slowly up a steep drainage that ended at a ridge, heavy timber on the far side. Burned timber. In the absence of wind, the ridge had acted as a fire line, and Union was simply going after the spots that had made it over. Spots were anything from a few square feet of ash to a vigorously flaming snag surrounded by an acre of black, but they all had to be put out. We were in sparse alpine fir now, Kelly Esterbrook said, and fire in alpine fir is very hard to fight. The trees blow sparks as they burn, igniting spot fires everywhere, and they grow in dense groves that provide critically high fuel loads when the fire reaches them. The flames climb up into the trees from the lower branches, which often reach to the ground, and make their way up into the densely packed crowns. A crown fire is particularly hard to stop because the fire moves from tree to tree without ever touching the ground. You can stop a crown fire only by cutting a lot of trees down, and after a certain point it becomes a case of cutting down the forest to save it.

The Union crew spread out in pairs to hit the spot fires and Miranda and I climbed on to the crest of the ridge. Below us, an entire river drainage of charred trees stretched away to the west. Occasionally a solitary tree torched down in the valley and sent up a tremendous plume of smoke. A temperature inversion had trapped the smoke low in the valley, and Miranda said that when it lifted—when the smoke started to rise—that meant that the air was turning over, and convection would invigorate the fire. At that point they would probably call in a substantial air attack in an effort to keep the fire contained on the west side of the ridge.

We followed the ridge, which climbed toward a peak that we had seen above camp. Below us on our right, the forest smoked. Below us on our left, hotshots worked the spot fires in small groups of three or four. An enormous Siller Brothers Skycrane helicopter with a bucket clattered up and down the valley, unleashing two thousand gallons of Smith Creek Lake water at a time. Four sawyers from the Smokey Bear crew dropped a flame-gutted ponderosa and then called in a water drop, and the Skycrane responded in minutes. Two thousand gallons, hitting a hillside from a hundred feet, is practically enough water to body surf downhill through the trees. When the water subsided, the sawyers bent over and began grubbing through the wet soil for embers.

Twenty minutes later, as they were finishing up, a message came over their radio from the fire camp: “Steve Shaeffer, Steve Shaeffer, Smokey Bear, your wife is in the hospital at this moment having a baby.” The sawyers grinned at one another; it was rare that overhead called with personal news that wasn’t bad. Shaeffer was farther up the ridge, hitting spots alongside the Negrito crew. He wasn’t able to go home, but at least he’d gotten the news.

At the top of the ridge we were met by the branch director, a strapping dark-haired man with a white-flecked beard and wind-burned face named Mike Rieser. He had fought fires since 1973 and was now fire control officer for the BLM Craig District in Colorado. We sat down and opened our lunches on a rock outcrop that overlooked the burning valley.

Rieser has personally known eight people who died on fires. “Wildland fire fighting has one of the highest incidences of fatalities and injury in the country,” he said. “In 1990 twenty-three people died, out of ten thousand active fire fighters. Six died on the convict crew at the Dude fire in Arizona, and that same week two were burned over in California. I saw a film of the first walk-through after the Dude fire; the heat varied so much that one shelter would be fine and the next one would have started to disintegrate.”

The Dude fire happened along the Mogollon Rim, on the north side of the Grand Canyon. It was a classic plume-dominated fire, but the topography of the rim served to intensify the downdrafts. Moments before the fire blew up, the Prescott Hotshots noticed a strange calm that often precedes a plume-dominated situation and radioed the overhead team that they were pulling out. The Perryville inmate crew and the Navajo Scouts crew were warned of the danger as well, but they were in exactly the wrong spot. The downdraft hit right in front of them and, funneled by the contours of the canyon rim, drove the fire straight toward them. Half the inmate crew escaped, as did the entire Navajo Scouts crew. The rest of the fire fighters dived into their shelters and waited for the flame front to pass over them.

From laboratory tests, researchers know that the adhesive that holds fire shelters together starts to melt at six hundred degrees Fahrenheit. That causes the fiberglass and aluminum layers to delaminate, in turn leading to rips and holes in the shelter. After the Dude fire, investigators found hard hats that had melted, leather gloves that had shrunk down to a couple of inches, and fire shelters that had begun to delaminate. Six men died, all from breathing superheated air. In all six cases they died because they had not deployed their shelters properly or because they had tried to leave them too quickly. Some of the survivors also suffered terrible burns, which suggest another possibility. The men who left the safety of their shelters may have done so because they thought they were dying inside them.

Mike Rieser and I sat on a rock outcrop eating our lunches and watching the fire. The sun was very hot and trees torched occasionally down in the valley. Every five minutes the Skycrane went by with a tremendous whump of rotors and released another two thousand gallons of Spring Creek Lake. Clouds began to move in above us, and Rieser said that they were called lenticular clouds and that the sculpted tops meant that the upper-level winds were over one hundred miles an hour. If those winds made it down to ground level, they would have a catastrophic effect on the fire. It was his job as branch director to watch out for things like lenticularis formations or castellanus or stage three cumulonimbus or anything else that might make the fire blow up. Two days earlier Rieser had pulled two entire divisions out of a canyon because he hadn’t liked the way the fire was behaving.

“We might not look like we’re doing much, but if you’re down in the valley digging, you don’t see what’s coming at you,” he said. “Air can rise off a fire, cool in the upper atmosphere, and rush back down. It superheats as it comes down and can overpower the wind field. That’s called a plume-dominated fire. It defies prediction. It’s what killed the people at the Dude fire; the ’shot crews recognized plume-dominated conditions, but the convict crew didn’t. No one could get word to them in time.”

Stage three cumulonimbus, thunderclouds, are a particular hazard. Not only do they introduce more lightning into the situation, but the air beneath their thirty-thousand-foot heads is extremely unstable. They can generate downward-moving winds of as much as one hundred miles an hour that hit the ground and spread out in a tremendous circle. The effect is to intensify the interaction among all three components of what is known as the fire triangle: fuel, oxygen, and heat. Ground winds spread dry, hot air through as yet unburned fuels, resulting in more fire and more heat, which in turn circulate more air, spreading the fire even faster. The result is a feedback loop that can be brought under control only if the relationships within the fire triangle break down—either by lowering the temperature of the fire or by depriving it of fuel. Water and retardant drops can bring the temperature of a fire down, and fire lines can deprive it of fuel. Otherwise, the fire spreads until the weather changes or there’s nothing left to burn.

In the Northern Rockies wildfire is usually started by lightning. Any lightning strike that reaches the ground can cause an explosion, but only lightning with a continuous current can start fires. In the Northern Rockies it has been estimated that one lightning stroke in twenty-five is a cloud-to-ground stroke capable of starting a fire. The inital bolt from cloud to ground moves relatively slowly, at one two-thousandths the speed of light, but it returns at one-tenth the speed of light and heats the gases inside it to fifty thousand degrees Fahrenheit. That much energy hitting a tree instantly raises it way past the ignition point, and often the tree just explodes. Flaming chunks of wood are hurled into the forest, and if the conditions are right for fire, the flames take hold.

Fire is a chemical reaction that releases energy in the form of light and heat. In the case of a wood fire, the energy was originally derived from the sun during photosynthesis and stored in the plant as cellulose and lignin. Heat—from a fire that is already burning or from a lightning strike—converts the cellulose and lignin into flammable gases, which are driven out of the wood and combined with oxygen in a process called rapid oxidation. At the base of any flame there is a clear band of superheated gases that have not yet ignited, a thin blue area of ignited gases, and a broad yellow band of incandescent carbon particles. The heat generated by this process continues to drive flammable gases out of more fuels, which burn and generate more heat. The heat also drives out moisture that might impede combustion. As long as there is sufficient air, fuel, and heat to ignite more fuel, the fire will keep advancing. As long as the fire keeps advancing, the fire triangle remains stable and continues to create the conditions necessary for fire.

A plume-dominated fire, more commonly known as a fire storm, is this same cycle writ large. In this case the three legs of the fire triangle not only provide the conditions for fire but amplify one another in an apocalyptic feedback loop—a “synergistic phenomenon of extreme burning characteristics,” as it is known. As in all fires, heat generates wind, which makes the fuels burn hotter, generating more wind. If a high fuel load is introduced into this loop, a convection cell of smoke and gases can be set in motion that overrides the local wind patterns. During World War II Allied bombers intentionally started fire storms in the German cities of Hamburg and Dresden; in that case the high fuel load was densely packed houses ignited by thousands of tons of ordinance. Once the convection engine has started, it is nearly impossible to stop. Entire stands of trees torch as one. Tornadoes twist through the interior of the storm. Superheated fuels appear to combust spontaneously in a phenomenon called “area ignition.” Such a fire can rip through well over one hundred thousand acres of timber in one day.

Another sort of apocalypse, equally destructive, is the running crown fire. Crown fires occur when flames climb so-called ladder fuels into the treetops and are swept along by high winds. In 1967 a running crown fire crossed the Idaho panhandle on a four-mile front that incinerated sixteen miles of timber in nine hours. The Sundance fire, as it was called, was calculated to have burned at rates of up to 22,500 British thermal units per square foot per second. By comparison, 500 Btu is the outer limit of what humans can control; 1,000 Btu describes potential fire storm conditions. The Sundance fire was estimated to release the energy equivalent of a twenty-kiloton Hiroshima type of bomb exploding every ten minutes.

Not all big fires are fire storms, of course, and not all fire storms are big. The Steep Creek blowup on the Lowman fire had the physical characteristics of a fire storm but was limited in area; the Foothills fire developed several convective columns over heavy timber but went on to become a wind-driven fire that ripped through two hundred thousand acres in two days, making it one of the biggest fire runs ever. In the Northern Rockies there are a host of winds that push fire: jet stream winds that drop down over mountainous areas; chinook winds that plunge downslope because of an air pressure differential; cold fronts that move in for twenty-four hours at a time; and, of course, unstable air associated with the fire itself. Unstable air rises and falls with the atmospheric conditions, whistling up canyons, sheering over ridges, bending around solitary trees or boulders to start whirls up to four thousand feet high. Any of those winds, in the wrong situation, could cause a blowup and kill people. That was why Mike Rieser was on a ridgetop watching the clouds rather than down in some canyon fighting the fire.

Hotshots have been known to complain that overhead—the men and women who risk other people’s lives—do not do enough. Not only that, but hotshots believe that many overhead have never really fought fire and therefore can’t be trusted to make life-and-death decisions. Sometimes that is true; there are the inevitable instances of ’shot crews simply saying, “No, we won’t go in that canyon,” or, “No, we won’t try to hold this ridge.” More often, however, the members of a command team, like Rieser, have worked their way up from grunt-hood to positions of authority over the course of years, if not decades. Rieser has fought fire for nearly twenty years and had two extremely close calls (that he told me about). Once he and his crew fell asleep after cutting line all night and were almost burned over; another time he was caught in a chaparral fire outside Los Angeles. Chaparral fires are extremely volatile because, invariably, the fuels are bone dry, the terrain is steep, and the winds are terrible: Santa Anas that hit seventy miles an hour for days on end. It was in 1979, and Rieser was on a type two crew that, he says, had violated just about every watch out rule in the book.

“We were backfiring off a road two-thirds the way up a ridge,” he said. “We couldn’t get a good burn because of a marine air intrusion. We double-shifted into the next day and got a dominating Santa Ana wind, and the fire just blew up. We were right in a canyon, it was acting like a natural chimney, and the flame front was on us in about ten minutes.”

In those ten minutes the crew managed to jump onto a tanker truck and make it to a marginally safe area at the intersection of two dirt roads, not big enough to qualify as a legitimate safe area but better than nothing. They parked the truck and crouched down between it and a road cut. The crew were so rattled that they were reading the Spanish side of their fire shelter instructions, not understanding a word. While the rest of the crew were trying to figure out their instructions, the fire went through.

“It was so loud that we couldn’t even shout to each other,” Rieser said. “It was not intolerably hot; the smoke was what was hard. We called in an air tanker and heard it make the drop about half a mile away; that was scary because we realized the smoke was so thick they didn’t even know where we were. It was an out-of-control situation, and our fate was in the hands of people who had made very questionable decisions. That was the turning point for me. I’ve been on worse fires, but I always pull the crews out before it gets bad. They say, ‘Aw, we could’ve held that,’ and then they watch it boil over.”

In 1871 a forest fire swept over the town of Peshtigo, Wisconsin, and killed more than fifteen hundred people. A fire in Chicago killed another three hundred people on the same day and is known as the Great Chicago Fire; the Peshtigo tragedy isn’t known as anything. Fires come in waves or complexes, and the Peshtigo and Chicago tragedies were part of a wide swath of fires that year that extended from Ohio to the High Plains. Since 1900 well over seven hundred people—it’s not known exactly how many—have died on wildfires in America, the vast majority of them men who were employed or had volunteered to fight the fires. The mass tragedies are mostly from the early days when there were no radios, no fire shelters, no aircraft, and no accurate weather forecasts. The next big fire complex after Peshtigo was the Big Blowup of 1910. Eighty-five men died battling the fires, some of them because they panicked and committed suicide after the fire lines were overrun. State troopers played taps over the caskets and buried them in mass graves in the hills. It was just this side of war.

In the end five million acres were burned in 1910, and flame-killed trees provided fuel for reburn cycles that lasted well into the 1930s. Reburns spread into healthy timber and ultimately redoubled the acreage that was lost. Faced with the possibility of a lumber shortage and consequently under tremendous political pressure, the Forest Service finally gave fire suppression top priority. Money was appropriated by Congress, crews were organized, lookout towers were built, timber companies constructed access roads into the mountains, telephones began to replace runners and mounted messengers. Fire fighting had finally entered, as one historian said, its heroic age.

The new approach showed results, but the West was still vulnerable to large-scale conflagration. The next big round of tactical changes came, not surprisingly, after the next big catastrophe, the Tillamook burn in 1933 that laid waste to half a billion board feet of Douglas fir in Oregon. Like the Big Blowup, Tillamook was merely the flagship of an armada of fires—Matilija, Selway, and others—that pummeled the West for three years. They prompted the Forest Service to adopt its famous 10:00 A.M. control policy, which meant that all fires were to be brought under control, if possible, by 10:00 the next morning. The policy would have been complete hubris were it not for a growing arsenal of fire-fighting tools: bulldozers that did the work of fifty men; airplanes that dropped thousands of gallons of retardant at a time; smoke jumpers who hit remote fires that would have taken men days to reach by foot. The idea was to get into the mountains fast and control the fires while they were still small; if a fire got away from the initial attack crews, thousands of lesser-trained men could be brought in to take over. As a tactic it made sense; as a public relations ploy it was unparalleled. Funding from Washington became effectively unlimited and remains so to this day.

As weather forecasting and communications improved, the loss of life declined, but mass tragedies still occurred. In 1937 a Civilian Conservation Corps crew of fourteen on the Blackwater fire were trapped between a spot fire and the main fire; they decided to turn and fight the spot rather than run, and they all died. In 1943 eleven marines died and seventy-two were injured when Santa Ana winds changed abruptly on the Hauser Creek fire in Southern California. In 1949 thirteen smoke jumpers and another man died in knee-high grass during a blowup on the Mann Gulch fire in Montana. In 1956 eleven convicts died on the Inaja fire in Southern California, and ten years later a hotshot crew lost twelve men in nearly identical circumstances on the Loop fire in the Angeles National Forest. A spot fire started at the bottom of a canyon, blew up unexpectedly, and ran twenty-two hundred feet of the canyon in less than a minute. That’s twenty-five miles an hour. The crew didn’t have a chance.

By the 1970s fire crews had portable fire shelters and Nomex clothing and could theoretically survive some burnovers. Fire shelters start to disintegrate at around six hundred degrees, though, and a fire run in heavy timber can hit temperatures three times that high. The only thing that will save a crew in the face of such an inferno is to get out of the way before it hits, and to that end researchers at the Intermountain Fire Science Laboratory in Missoula, Montana, have developed mathematical models that predict—given certain fuel conditions, terrain type, and meteorological conditions—what a fire will do. These models have been programmed into computers and can be used in conjunction with satellite data to project fire growth on any fire anywhere in the United States. The incident command team on a fire can punch its location into a computer, along with topographical and meteorological information, and receive very specific information about likely fire behavior the following day: which canyons will burn out; which ridgelines will hold.

Still, no amount of computing power can predict exactly what a fire will do.

“Nothing we’re doing today is more important than a human life,” one incident commander said at a 6:00 A.M. briefing.

That sentence—more than fire behavior models, lightweight fire shelters, or advanced meteorology—explains why people no longer die in the terrible numbers they used to on wildfires in the United States.

In the middle of the afternoon we moved down off the ridgetops toward H-6. Helicopters were to lift us and the three crews down to the fire camp for the night. I was in no hurry to leave the mountains, but the Union, Negrito, and Smokey Bear crews were being replaced by type two outfits that hadn’t spent the last two nights spiked out. We clustered around the helispot and then piled onto the helicopters with our line packs and were shuttled back to camp.

That night I heard that a fire had just leaped the huge Salmon River Canyon up by McCall; a fire had started north of the river and a backburn had been set by dropping flaming Ping-Pong balls from a helicopter. (The Ping-Pong balls are filled with potassium permanganate and ignited by antifreeze. A needle injects the antifreeze at the moment of release and they ignite after thirty seconds, usually after they hit the ground. The helicopter pilot had circled to check his work, and one of the balls had accidentally released over the south side of the river. That was all it took. Within hours the new fire was fifty acres in size and crowning through heavy timber.

I drove up to McCall the next morning to see if I could get on the fire, but by then it was so far out of control that no one was allowed near it. Crews were cutting indirect line miles ahead of the fire and cold-trailing their way southward from the river, but that was it. The town of Riggins was filled with smoke; there was a mushroom cloud of smoke pumping out of the canyon; the airfield at McCall was crammed with reserve helicopters. But I wanted to see real flame. I slept on a sandbar along the North Folk of the Salmon River and drove back to Boise the next day. There was another fire north of town, I was told. It was racing through the dead-brown hills, and hotshots were getting pulled off Foothills to deal with it. It was bad, and it was moving fast.

There’s a little river that runs through Boise, and from the cafés that line the river walk, I could look up and watch the mountains burn. A big head of smoke was pumping out of the hills to the north, and retardant-streaked air tankers were making nonstop runs to and from the tanker base at the Boise airport, south of town. According to dispatch, bulldozers were trying to save a subdivision off Highway 21, wind-driven flames were racing through the terrifically dry sagebrush and cheatgrass around Lucky Peak Reservoir, and fifty crews left over from the Foothills fire were waiting to go in. I picked up a pass for the roadblock and headed north up Highway 21, toward the smoke.

Manpower was so short that the roadblock was guarded by a middle-aged couple in lawn chairs. I showed them my letter, and they waved me through. Soon I was alone on the dirt road that led into the burn zone. A line of flame hung like a necklace along the parched flanks of the hills. Smoke had turned the sunset blood-red. After three or four miles there was a hand-written cardboard sign that read: AREA CLEARED @ 19:30 HOURS 9/2—U.S.F.S. Just beyond that was the fire. It had reached the road and was swirling around a utility line that continued on up into the hills. I stopped the car and got out, completely alone with the fire and the mountains and the huge dead sky. Ten-foot tongues of flame licked the guardrails and shot into the sky. The vegetation died loudly, as if in pain, popping and exploding in the thickening dusk.

I took a few photographs and then went back to my car. The fire was about to jump the road. It would eventually move up into some timber and end up torching over thirteen thousand acres. A house would burn down. The beautiful Leonard ranch would be saved—barely—by ground and dozer crews backed up by massive air attack. It was called the Dunnigan Creek fire. It was one of roughly one hundred thousand wildfires during the summer of 1992, and if you ask a hotshot if he’s ever heard of it, chances are he’ll say no. Rain put it out after a couple of days.

This essay was written in 1992, and since then there have been many changes in the way wildfire is fought in the United States. I wanted to alter the original work as little as possible, so it should be noted that—among other changes—the Boise Interagency Fire Center is now called the National Interagency Fire Center; fire fighters are paid more than $8.50 an hour; and portable computers are in common use for predicting fire behavior.

Most significant, however, the fatality statistics have changed. When I was in Idaho, it had been twenty-six years since more than half a crew had died in a single incident—the Loop fire in southern California, where twelve hotshots were overrun in a matter of minutes. Tragically, in 1994, fourteen hotshots and smoke jumpers were overrun in similar circumstances on the South Canyon fire outside Glenwood Springs, Colorado. The South Canyon fire is now the deadliest fire since 1937. My account of that incident appears as “Blowup,” the second essay in this collection.

Many hotshots I spoke with attributed the increasing danger of their job to severe drought conditions in the northern Rockies, as well as to decades of rigorous fire suppression. Both have contributed to a huge buildup of dead fuel in our nation’s forests—fuel that ordinarily would have been cleared out by the small fires that regularly flare up in an unmanaged ecosystem. A disastrous fire season was inevitable, and in 2000 it finally happened. Eighty-five thousand wildfires burned nearly seven million acres across the United States. Sixteen people died, and fire suppression cost over one billion dollars.

It was the worst season ever. With the western drought continuing unabated and huge amounts of deadwood still choking many forests, fire behavior experts don’t expect conditions to get better anytime soon.