Fire today ManagementVolume 65 • No. 2 • Spring 2005

LLARGEARGE FFIRESIRES OFOF 2002—P2002—PARTART 22

United States Department of Agriculture Forest Service Erratum In Fire Management Today volume 64(4), the article "A New Tool for Mopup and Other Fire Management Tasks" by Bill Gray shows incorrect telephone and fax numbers on page 47. The correct numbers are 210-614-4080 (tel.) and 210-614-0347 (fax).

Fire Management Today is published by the Forest Service of the U.S. Department of Agriculture, Washington, DC. The Secretary of Agriculture has determined that the publication of this periodical is necessary in the transaction of the pub- lic business required by law of this Department.

Fire Management Today is for sale by the Superintendent of Documents, U.S. Government Printing Office, at: Internet: bookstore.gpo.gov Phone: 202-512-1800 Fax: 202-512-2250 Mail: Stop SSOP, Washington, DC 20402-0001

Fire Management Today is available on the World Wide Web at http://www.fs.fed.us/fire/fmt/index.html Mike Johanns, Secretary Melissa Frey U.S. Department of Agriculture General Manager

Dale Bosworth, Chief Robert H. “Hutch” Brown, Ph.D. Forest Service Managing Editor

Tom Harbour, Director Madelyn Dillon Fire and Aviation Management Editor

Delvin R. Bunton Issue Coordinator

The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, sex, religion, age, disability, political beliefs, sexual orientation, or marital or family status. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communica- tion of program information (Braille, large print, audiotape, etc.) should contact USDA’s TARGET Center at (202) 720- 2600 (voice and TDD).

To file a complaint of discrimination, write USDA, Director, Office of Civil Rights, Room 326-W, Whitten Building, 1400 Independence Avenue, SW, Washington, DC 20250-9410 or call (202) 720-5964 (voice and TDD). USDA is an equal oppor- tunity provider and employer.

Disclaimer: The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader. Such use does not constitute an official endorsement of any product or service by the U.S. Department of Agriculture. Individual authors are responsible for the technical accuracy of the material presented in Fire Management Today. Fire Management today Volume 65 • No. 2 • Spring 2005

On the Cover: CONTENTS Monster in the Woods: The Biscuit Fire ...... 4 Beth Quinn How Did Prefire Treatments Affect the Biscuit Fire? . . . . 18 Crystal Raymond and and David L. Peterson Snow Camp Lookout: Remembering a Biscuit Fire Casualty ...... 23 Jan Robbins Prediction Errors in Wildland Fire Situation Analyses ...... 25 Tree silhouetted against Geoffrey H. Donovan and Peter Noordijk the massive Biscuit Fire as Fire in Our Mountains— it makes its way through and Mountains in Our Rivers ...... 28 Oregon’s Siskiyou National Thomas W. Swetnam Forest toward the Illinois Valley. Photo: Thomas Iraci, Treatment Success on the Rodeo–Chediski Fire ...... 30 USDA Forest Service, Paul Keller Portland, OR, 2002. Operation Success: Columbia Space Shuttle Recovery ...... 32 Traci Weaver Shuttle Recovery: Largest GIS Emergency Response to Date ...... 35 Traci Weaver Using Satellite Imagery for Burned Area Emergency Response ...... 37

The FIRE 21 symbol (shown below and on the Andrew Orleman, Jess Clark, Annette Parsons, and Keith Lannom cover) stands for the safe and effective use of wild- land fire, now and throughout the 21st century. Its Satellite Mapping of Wildland Fire Activity ...... 40 shape represents the fire triangle (oxygen, heat, and fuel). The three outer red triangles represent Keith Lannom, Brad Quayle, and Mark Finco the basic functions of wildland fire organizations (planning, operations, and aviation management), and the three critical aspects of wildland fire man- agement (prevention, suppression, and prescrip- SHORT FEATURES tion). The black interior represents land affected by fire; the emerging green points symbolize the growth, restoration, and sustainability associated Websites on Fire ...... 17 with fire-adapted ecosystems. The flame represents fire itself as an ever-present force in nature. For more information on FIRE 21 and the science, Guidelines for Contributors ...... 42 research, and innovative thinking behind it, contact Mike Apicello, National Interagency Fire Center, Annual Photo Contest ...... 43 208-387-5460.

Firefighter and public safety is our first priority.

Volume 65 • No. 2 • Spring 2005 3 MONSTER IN THE WOODS: THE BISCUIT FIRE*

Beth Quinn

vast plume of thick gray Greg Gilpin, a fire manager for the blaze had fallen back, leaving no smoke billowed 30,000 feet Oregon Department of Forestry, defenses between the flames and the A (9,000 m) into the sky over could sense the fear in the 1,500 17,000 people living in the Illinois southwestern Oregon, dwarfing people gathered before him at Valley towns of Selma, Kerby, Cave ridgetops on the Siskiyou National the Illinois Valley High School on Junction, and O’Brien. Forest. Six more mushroom clouds Sunday, July 28. His job was protect- rose above nearby ridges like volca- ing their homes and lives, and he Gilpin stepped onto the polished nic eruptions. Trees 150 feet (45 m) knew that the situation was even wooden floor, stood under a basket- tall torched like matchsticks. worse than it looked. ball hoop and spoke into a micro- phone. “There is a very good chance Valley in Danger that this fire is going to reach the Nearly 7,000 of the valley floor,” he said. “It is so big By mid-July 2002, the Florence Fire Nation’s best firefighters and so awesome, there is absolutely (later known as the Biscuit Fire**) nothing you can do to stop this fire.” was burning with a destructive defied the odds and power rarely seen. Already the big- held off an inferno that Yet the worst never came to pass. gest Oregon blaze in more than a threatened hundreds Nearly 7,000 of the Nation’s best century, it was poised to roar down of square miles and firefighters defied the odds and held into the heavily populated Illinois thousands of homes. off an inferno that threatened hun- Valley. Two weeks of ceaseless strug- dreds of square miles and thousands gle had done almost nothing to stop of homes. The Biscuit Fire, which it, and firefighters put the chances On the other side of those ridges, eventually eclipsed the infamous of losing towns in the valley at 75 Tillamook Burn as Oregon’s sig- percent. just out of sight, a 20-mile (32-km) wall of fire was moving through a nature fire, offers lessons for every Beth Quinn is a southern Oregon correspon- wilderness of tinder-dry trees and State in the West. dent for The Oregonian, Portland, OR. brush. Every firefighter facing the Ignition Southwestern Oregon baked under a blistering heat wave. Sunrise on July 13 marked the 53rd day since the last rain, and the weather forecast that day called for a high of 105 °F (41 °C). The forests in the Illinois Valley were as parched as kiln-dried lumber.

* The article is based on a series that appeared in The Oregonian (Portland, OR) on November 2, 3, and 4, 2002. Reporter Alex Pulaski and researchers Lynne Palombo and Kathleen Blythe of The Oregonian staff contributed to the series.

**Florence became Biscuit to avoid public confusion Biscuit Fire on August 16, 2002. A smoke plume rises from the vicinity of Snow Camp about the location of the fire. Some people wondered Lookout as the northwest flank of the fire creeps down into the Lawson Creek drainage. In whether the fire was near Florence, a community adjacent the foreground are Oak Flat and the Illinois River. Photo: Gary Percy, Siskiyou National to the popular Oregon Dunes National Recreation Area Forest, Gold Beach, OR, 2002. on the Siuslaw National Forest. Consistent with normal nomenclature for wildfires, this article refers to the fire for the most part as Florence.

Fire Management Today 4 Conditions were perfect for the dry elsewhere in Oregon and the region. (40,000 ha) in Oregon alone. The lightning that strikes when heat None would be available for at least Northwest Interagency Coordination evaporates a thunderstorm’s rain- 48 hours. Del Monte was on his own. Center in Portland weighed the des- drops before they reach the ground. perate pleas from local fire manag- One-third of forest fires begin with a Initial Attack ers. single explosion of dry lightning. A trail passed near one of the fires, and Del Monte sent two local fire Federal rules set strict guidelines Firespotters eyeballed the woods crews on a 7-mile (11-km) hike. for making such decisions. The first from glass-walled lookouts on five Their orders were to contain the fire priority is human life, then towns different mountaintops. At midday, and, if possible, put it out. and historically significant cul- thunderheads rose above the moun- tural resources, such as American tains. The storm erupted just after 2 Del Monte climbed aboard a heli- Indian pictographs or archeologi- p.m., unleashing a fusillade of thun- copter for a short flight to the other cal sites. Farther down the list are derbolts on the thickly forested land fire. The view across the sea of vacation homes and, finally, timber along the California–Oregon border. green ridges was ominous: Two fires in uninhabited forests such as the Sensors recorded 581 downstrikes were evident where the day before Kalmiopsis. Del Monte’s trees would in Jackson and Josephine Counties, there had been just one. Both were have to wait their turn. with 23 thunderbolts blasting onto on steep and rocky slopes, where the Siskiyou National Forest. Not a glowing red edge of fire moved Federal fire coordinators informed a drop of water slaked the thirsty through Douglas-fir and white fir. the Siskiyou fire managers that Earth. they were second in line for a type Del Monte sent a team with two 2 team, behind crews fighting One lightning bolt struck near bulldozers to widen the rough jeep the Trimbly Creek Fire in eastern Florence Creek, touching off a small tracks near the two fires. He knew Oregon, where a mile-wide flame fire that hid beneath heavy brush from the beginning that the effort front threatened six homes. and thick trees in the Kalmiopsis was likely to be futile, at least in the Wilderness. Firefighters call them short run. The fires would almost The situation turned grimmer on “sleepers,” fires that can creep certainly be out of control before July 15, when a resident reported unseen for days before exploding. heavy equipment could use the yet another fire in the rough back- tracks. country. And then, that afternoon, As the storm abated, a Siskiyou a reconnaissance flight spotted one National Forest reconnaissance plane more fire, this one on a south-facing took flight to search for fire. The slope near Florence Creek, 27 miles spotter and pilot quickly found two Local crews could fight (43 km) north of the other blazes. other blazes pumping out smoke vis- small fires, but not three Florence had finally reared its flam- ible from the air. blazes that were miles ing head. apart, each burning Bob Del Monte, assistant fire man- through roadless areas. First View of Florence agement officer for the Siskiyou Del Monte had a single fire crew National Forest, immediately rec- in reserve, which meant he could ognized the danger. The isolated The Siskiyou fire managers quickly fight only one of the other four fires Kalmiopsis Wilderness was a fortress agreed that they were overwhelmed. burning on the Siskiyou. He chose for fire. A vertical landscape of sharp Local crews could fight small fires, Florence because it posed the great- ridges and plunging canyons, thread- but not three blazes that were miles est danger to a populated area. It was ed by rushing water and scarred by apart, each burning through road- only 6 miles (10 km) from the ham- old burns, the wilderness contained less areas. They asked for a type 2 let of Oak Flat, a 480-acre (190-ha) only a handful of roads—rough team, but they were hardly the only island of private land on the Siskiyou tracks that demand four-wheel drive. ones making such a request. This National Forest, including 2 homes was the second-worst western fire and 12 summer cabins. Del Monte Del Monte asked for smokejump- season in 50 years. Across the Pacific had confidence in the crew boss, ers, but dispatchers at the Redmond Northwest, fires were erupting on a Paul Hiebert, who had experience as Air Center turned him down. The broad front. A dozen major fires were a hotshot. smokejumpers were fighting fires burning on nearly 100,000 acres

Volume 65 • No. 2 • Spring 2005 5 The crew set out for the fire at 7:45 p.m. on July 15. After hiking 4 miles The watchword among fire managers was “safety (6 km) into the darkness, the crew first,” and they wouldn’t risk lives to protect an bedded down beside the trail. They uninhabited forest. couldn’t see, smell, or hear the fire. Early next morning, they set out again, catching their first sight of forest. He plotted an escape route to ha). At 8 p.m., Del Monte told the Florence at 8:15 a.m., just as they use if the winds shifted and the fire crew to pull out. rounded a bend along the Illinois wheeled around his crew. River. Florence had won the first round. Florence was well dug in. The flames Although two crews had contained It was already an impressive speci- were about a quarter mile (0.6 km) one of the Siskiyou fires, the others men. The flames had spread across from the Illinois River, a steep uphill were burning out of control. 50 acres (20 ha) and were tearing hike. The only possible route for through 15-foot (5-m) brush and a retreat was through a deep, boulder- Dwindling Options tangle of dead trees along the can- strewn creek bed. If the crew had to Florence doubled again in the day yon wall. Embers wafted by upslope run for it, Hiebert thought, someone after Hiebert’s retreat and was send- winds carried spot fires hundreds of could fall and break a leg. ing columns of smoke 10,000 feet feet up the ridge. Fingers of flame (3,000 m) into the air. If not stopped crawled toward the river on fallen Hiebert went over the options again now, it could burn until late fall, snags. and again. His crew couldn’t get scorching the forest through the dry around the fire, which was spread- days of August and September. Hiebert’s heart sank. The fire had ing in all directions. He radioed the covered 6 acres (2.4 ha) when the bad news to Del Monte. As Hiebert Del Monte studied the map, search- team was dispatched and had grown waited for an answer, the crew set to ing for a strategy that would keep 10 times as large in just a day. work clearing brush along Florence Florence contained in the wilderness Hiebert had hoped the fire would be Creek. until help could arrive. The terrain high on the ridge, but instead it was offered some outer boundaries for burning close to Florence Creek and As they worked, the fire grew— the blaze. To the north, a gravel bar- a smaller creek nearby. That made it jumping across one of the creeks rier, Bald Mountain Road, served as a even more dangerous. and thundering up the ridge with a fire break. To the south, the Illinois shriek that reminded Hiebert of a jet River provided what might be a suf- Sizeup engine’s roar. By midafternoon, the ficient natural barrier. fire had swelled to 300 acres (120 Hiebert left his crew beside the river and sized up the fire. The standard tactic for taming such a blaze was a direct attack. Once the fire was con- tained, Hiebert could run a hose to the river and use a portable pump to douse the flames. Or he could call in an airstrike of water or retardant.

As he calculated the angles of possi- ble attack, Hiebert thought about his crew’s safety. No one fighting wild- fires had forgotten the deaths of four firefighters on the Thirtymile Fire during the previous summer. The watchword among fire managers was “safety first,” and Hiebert wouldn’t risk lives to protect an uninhabited Biscuit Fire nearing the Illinois River and threatening small communities. Photo: Tom Iraci, USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002.

Fire Management Today 6 Yet every direction posed serious for reinforcements, calling for a type attack firelines. Again and again, the problems. Winds blowing every after- 1 team. southwestern Oregon crews had to noon along the Illinois River canyon fall way back from the leading edge had already nudged Florence east But 14 of the 16 type 1 teams were of the fire, sacrificing huge stands of toward Oak Flat. If the fire jumped already committed elsewhere, timber in exchange for a safe place the narrow river and headed south, including 4 on other Oregon fires. to make a stand. it would storm into the Kalmiopsis The coordination center in Portland and its enormous amounts of fuel turned him down. The Siskiyou blaz- Joki deployed a second team of fire- and lack of natural barriers. es, the fire coordinators said, posed fighters along Bald Mountain Road. no imminent threat to human life or Their mission was to keep the fire, Air tankers could push back a fire of communities. Instead, Portland sent growing slowly northward, penned this size. But the 10 Oregon tank- a second type 2 team to the Illinois in the wilderness. But Florence had ers were already protecting homes Valley to battle the southern fires: its own strategy. As the crews dug in threatened by fires in eastern and Biscuit 1, Biscuit 2, and Sourdough. with their bulldozers and chain saws, central Oregon. The danger to Oak the fire suddenly turned east, gallop- Flat wasn’t enough to justify a That day, the three fires merged into ing along the road toward Oak Flat. change in plans. a single blaze. The fire managers named it Sour Biscuit. Threat to Oak Flat On July 21, a week after the Firefighters raced ahead to set up Northwest Interagency Coordination Joki turned back to the bigger, more a new blocking position that would Center approved the request, a type dangerous threat: the implacable, protect the 14 houses in Oak Flat, 2 team arrived. Glenn Joki, a fire capricious Florence blaze. now just 4 miles (6.4 km) from the manager with 37 years of experi- flame front. Joki ordered his crews ence, led the team, which worked to prepare the ground for a huge out of Arizona. The firefighters were Fire managers have burnout, as large as 1,920 acres just off the giant Rodeo–Chedeski acknowledged that they (780 ha). The firefighters planned Fire, a conflagration that made the were stretched thin by to blacken 3 square miles (8 km2) tiny Arizona town of Show Low the of forest between Florence and the national focus for the summer’s an outbreak of wildfires hamlet. wildfires. across 11 States and missed an early Joki wanted to know what route Joki flew over the fires. He didn’t chance to put out the civilians and firefighters could take want to see flames; he wanted to see if the fire turned unexpectedly, and country. From the air, the forest’s Florence blaze. he scouted the ground himself. web of creeks was invisible, running The Illinois River Road, he saw, through the bottom of deep and was surprisingly rugged. It was a narrow canyons. In the distance, Digging In narrow, partially paved track with Joki could see roads through the Joki launched an ambitious plan. He hairpin turns, steep grades along a forest, but only Bald Mountain Road sent half of his crews to build a 5- cliff above the river, and washouts. A lay close to the Florence Fire. Joki mile (8-km) fireline along Florence’s retreat along this road would be slow looked for a place where his fire- eastern flank. They set to work more and dangerous. fighters could dig in and know the than 4 miles (6.4 km) northeast of fire wouldn’t get around them. He the fire, a decision that consigned The scene at Oak Flat was even more found few good options. thousands of additional acres to the discouraging. Joki’s trained eyes saw flames. Joki felt he had little choice. homes nestled deep in the woods, The three other fires burned toward No place closer to the fire offered an with trees directly overhead. Many one another through the south- adequate margin of safety. sat at the end of driveways so over- ern end of the Siskiyou National grown they looked like tunnels. His Forest, miles away from Florence. Florence was teaching a lesson: firefighters attempted to clear brush Joki realized that he couldn’t battle With today’s huge fuel loads, fires and build a perimeter around the big blazes so far apart. He and the in rough country such as the houses, reducing the chance that Siskiyou fire managers asked again Kalmiopsis can easily overrun direct- stray embers could set them afire.

Volume 65 • No. 2 • Spring 2005 7 As Joki’s crews worked in Oak Flat, A plume of brilliant white smoke Florence unleashed plume after Florence once again changed direc- and steam spiraled above a nearby plume. After leaping 100 feet (30 tions, moving south toward the ridge, forming the thunderhead that m) across the Illinois River, the fire Illinois River, posing a whole new set signals unpredictable, explosive fire churned toward the valley towns, its of dangers. The fire could follow the danger. The blaze that had broken flames whipped by 60-mile-per-hour river into Oak Flat, forcing evacu- out just 11 days earlier was now (100-km/h) winds from Oregon’s ation. And Joki began to wonder: spawning a tornado of fire, fueled by inland high desert. The Chetco What if Florence leapt across the dense forest and fanned by blister- winds, named for the river valley 100-foot-wide (30-m-wide) Illinois ing winds. Firefighters immediately they follow to the sea, usually arrive River, the only natural barrier left recognized the danger. When the in late October. But, for the first between the fire and the people liv- plume topped 25,000 feet (7,600 m), time in anyone’s memory, they were ing in the Illinois Valley? it would collapse, spewing embers blasting through July afternoons, miles from the fire’s leading edge. with temperatures topping 100 °F The answer came the next day. On (38 °C), working like a bellows on a July 24 at 4 p.m., Joki took a call as The firefighters told Oak Flat fire that already had ample fuel and he left a meeting of fire managers residents to flee. But one stubborn momentum. in Medford. Florence had jumped couple refused to leave their farm the river and was tearing up a slope undefended, so firefighters handed “You have to recognize that it’s toward the next mountaintop. Three them a pair of fire shelters. If the beyond your control,” said Joki. helicopters were dousing the fire flames overwhelm you, they said, “On those days, the dragon wins.” with water, to no effect. “There it run into the blackened meadow and But Joki could not let this dragon goes,” Joki told his colleagues. crawl inside. The firefighters lighted prevail. His firefighters, 663 strong, a burnout to clear the threatened formed the last line of defense for The Better Part of Valor ground near homes. They waited just the Illinois Valley. It was a daunting On July 24, the fire crews that had long enough to see the brush burst prospect. been frantically clearing brush into flames, then they jumped into around the hamlet of Oak Flat their rigs and raced out of Oak Flat. watched in horror as the Florence With today’s huge fuel Fire revealed its new, nightmarish Out of Control loads, fires in rough powers. Over the next week, the firefight- country such as the ers retreated again and again as Kalmiopsis Wilderness can easily overrun direct- attack firelines.

A fire hot enough to create plumes is a capricious beast. Such wildfires don’t follow the usual paths along river bottoms or up slopes. And no matter which direction a plume takes, the fire races at speeds of up to 10 miles per hour (16 km/h), far too fast for firefighters to outrun.

The plumes were awe inspiring, frightening even to men who spent their lives around fire. They made a sound unlike anything else on Earth, an ear-splitting shriek caused by Sikorsky helicopter taking on retardant at the confluence of the Illinois and Rogue Rivers winds uprooting trees, tossing logs, while battling the Biscuit Fire on August 20, 2002. Photo: Gary Percy, Siskiyou National and filling the air with burning mis- Forest, Gold Beach, OR, 2002.

Fire Management Today 8 That night, Joki faced about 300 Illinois Valley residents at the Josephine County Building in Cave Junction. Joki stood before the crowd and delivered the grim news. The first of Florence’s plumes had collapsed that day, and the fire had taken off in every direction, chewing up 10,000 additional acres (4,000 ha) in the Kalmiopsis Wilderness. That brought the fire to 15,932 acres (6,373 ha) and put it on a course directly toward the Illinois Valley.

“Ten thousand acres in one day?” one woman asked.

Joki struggled to explain. The ter- ror of an uncontrolled wildfire was unimaginable for most people, even those whose backyards were a 1.2 million-acre (490,000-ha) national forest. Little more than a month before, Joki had watched residents in Arizona absorb similar terrifying news, even as the Rodeo–Chedeski Fire was sending 426 houses up in smoke. For the public, wildfire didn’t seem real until the blowtorch headed for their homes, their subdivisions, or their towns.

Joki had more worrisome predictions for local residents. The meteorolo- gists on his team warned of more hot, dry, blustery weather, which could cause additional explosive fire Biscuit Fire glowing eerily as it makes its way toward the Illinois River. Photo: Tom Iraci, plumes. Gilpin, the State fire manag- USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002 er, warned that the fire could return to areas already burned, so residents siles—the cones of ponderosa pines couple who had refused to leave. of Oak Flat wouldn’t be allowed back and the limbs of burning trees. to their homes. Joki saw it as a victory. He walked Illinois Valley Imperiled among the fire crews scattered The weather cooperated on July 26, On July 25, the day after they had throughout the hamlet, surveying with a layer of warm air trapping the fled Oak Flat, firefighters returned the scorched homes near the river cooler air below, temporarily slowing to see what had survived. Their des- and the blackened canyon walls ris- the fire. But Florence roared back peration move had worked, up to ing above. But Oak Flat was not yet the next day. As the day warmed and a point. The fire reached Oak Flat, safe. Fire could still burn through the winds began to blow, it spun out incinerating two cabins, a barn, a the forest from the north. The crews another plume and bolted south, Quonset hut, and a shed. Ten cabins set to work again, dousing hotspots toward Oak Flat. and two homes survived, as did the and building a new fireline.

Volume 65 • No. 2 • Spring 2005 9 Joki’s firefighters again set off a Highway 5 travelers descending into outside the high school and watched burnout. But instead of retreat- Grants Pass, 26 miles (42 km) away. the Florence Fire advance to the ing up the Illinois River Road, they In 37 firefighting seasons, Joki had ridgetops, torching 150-foot (45-m) remained in Oak Flat in a new safety never seen a fire like Florence. Every trees. zone—a burned-over meadow—and wildfire in his experience would slow watched the fire erupt around them. at night, as the temperature fell. Inside the gym, Gilpin scanned the With them were 17 civilians, includ- But this one had so much heat and worried faces. Gilpin knew these ing the same couple who had refused momentum that it cranked right woods, knew just how volatile these to leave their farm undefended when through to morning, making tre- forests were, and yet in 25 seasons Florence had roared into Oak Flat mendous runs after dark and burn- he had never seen a fire run this three days earlier. They were joined ing intensely 24 hours a day. hard at this many people. by 15 visiting family members and friends. “If you live anywhere in the Illinois When the plume topped Valley, you need to start thinking Florence assaulted Oak Flat with 25,000 feet, it would about what you’ll do if you have to 100-foot (30-m) flames, searing heat, evacuate,” Gilpin said. “We do need and powerful winds that launched collapse, spewing to be prepared for the possibility of spot fires 2 miles (3.2 km) ahead of embers miles from the fire on the valley floor in the next 24 the flame front. As the woods burned fire’s leading edge. to 36 hours.” around them, Joki’s firefighters and the civilians saw several more out- It was time to gather irreplaceable buildings go up in smoke. And they The weather forecast was nightmar- items, Gilpin told the stunned resi- heard deafening explosions erupt ish: hot days, cloudless skies, and dents, and to pack them into cars as flames detonated propane tanks bone-dry winds across the ridgetops. and trucks, then to turn those vehi- stashed as far away from homes as cles around and to park them facing possible. On July 28, a hot blue-sky Sunday out of their driveways. To Gilpin, the afternoon, Florence kicked up seven unthinkable was now a stunningly Fire managers wondered whether billowing plumes. Each collapsed, real possibility. The unseasonable Oak Flat was a harbinger of the fate spitting flame 2 miles (3.2 km) Chetco winds out of the northeast facing the entire Illinois Valley. They ahead of the fire’s core. Through the had fanned the Florence Fire into a told residents of 30 houses and of afternoon, each rose and fell 3 times, 20-mile (32-km) flame front. If and the historic McCaleb Ranch along for a total of 21 separate incidents. when the normal west winds of July the Illinois River Road to abandon returned, the long wall of fire would their homes, and they warned anoth- That day, Florence chewed through blow into the Illinois Valley. er 742 homeowners near Selma to 45,000 acres (18,000 ha). The fire prepare for evacuation within 48 traveled nearly 7 miles (11 km) The Northwest Interagency hours of official notice. By day’s north and 9 miles (14 km) south. Coordination Center in Portland end, the Florence Fire had grown to When the day was done, the McCaleb had reached the same conclusion. 23,270 acres (9,308 ha). Ranch and a remote forest cabin had Citing the flame front paralleling been incinerated. U.S. Route 199, they said there was In Eagle Creek, 250 miles (400 km) a 75-percent chance that the fire north, Mike Lohrey’s cell phone That evening, 1,500 frightened would reach one or all of the Illinois rang. On the line was Bob Del people streamed into the parking Valley’s four towns in the next 2 to Monte, under normal circumstances lot at Illinois Valley High School. 5 days. Six days after Siskiyou fire a member of Lohrey’s type 1 team. “I Throughout the afternoon, they’d managers first asked for a type 1 think you’re coming down here,” Del watched the plumes climb and col- team, the Portland center forwarded Monte told Lohrey. “This thing has lapse in the clear blue skies. Along the request to the Multiagency gone nuts.” a 20-mile (32-km) stretch of U.S. Coordinating (MAC) Group in Boise, Route 199, a two-lane road that which assigns the Nation’s 16 type 1 Monster on the March bisects the Illinois Valley, the view teams. The MAC Group didn’t hesi- Late that night, Florence’s orange to the west was an ominous string tate. Lives were threatened. It was glow was visible to Interstate of plumes towering over the valley. time to call in the best firefighters Knots of anxious people gathered the Nation could muster.

Fire Management Today 10 Days of Fear fire. They removed brush, limbs, the story: The Florence Fire had and trees where possible and plotted reached 141,650 acres (56,660 ha) The Florence Fire’s Sunday ram- the location of each with a satellite and the Sour Biscuit 33,287 acres page persuaded many Illinois Valley global positioning system. (13,315 ha). residents to evacuate even before an official notice. A steady stream of Local officials recognized that they The gym was eerily dark, the lights packed vans, campers, utility trail- were confronting a force of nature left off in hopes of keeping the air ers—anything with wheels—headed that could blaze any path it chose. cooler. out of the Illinois Valley. “For everything we’ve tried to do,” said Illinois Valley Fire District Chief Gilpin stood before the crowd Heading into the remote valley was Kyle Kirchner, “this fire has reared arrayed across a wall of bleach- another convoy of vehicles—bulldoz- up and kicked us in the face.” ers, turning a slow half-circle as he ers on huge flatbeds, red and yel- described a 30-mile (48-km) make- low fire engines from almost three The unseasonable Chetco winds shift rescue line to protect their dozen Oregon communities, vans died down, a significant break for homes, a line that remained days and school buses carrying firefight- the firefighters. Florence, however, from completion. “I want you to ing crews. But for every Illinois continued its march. And the Sour be very aware that when the wind Valley resident choosing to evacuate, Biscuit blaze that was born in the switches, even if we have a line on many more elected to stay—at least same lightning strike as Florence the east side of the fire,” Gilpin said, temporarily—and a few even talked had picked up momentum as well “I can’t guarantee that we’re going about making a private stand against and now burned just 3 miles (4.8 to hold that line. If this fire moves the approaching flames. km) from the larger fire. Soon, they to the valley floor, I cannot guaran- would join. tee that we can stop it on the valley At the Selma Market, information floor. officers huddled with worried resi- dents, tracing the fire’s perimeter on “We’re looking at a fire that effec- 3-by-5-foot (0.9-by-1.5 m) maps and Florence assaulted tively at this point in time is uncon- advising people to leave immediately Oak Flat with 100-foot trolled. We’re looking at the possibil- rather than risk an enormous traf- flames, searing heat, ity of this fire burning to the Rogue fic jam—like “getting caught in a and powerful winds that River,” Gilpin said. “We’re looking at grade B movie,” they called it—if the the possibility of this fire jumping evacuation order came. launched spot fires 2 miles ahead of the the Rogue River. And we’re looking at the possibility that this fire could To the south, things looked dicey. A flame front. go to 500,000 acres [200,000 ha].” separate fire near Gasquet, CA, had closed U.S. Route 199 south, leaving Everyone in the Illinois Valley open only two roads leading from Thirty Minutes To should be prepared to evacuate their the Illinois Valley to safety: U.S. Evacuate homes within 30 minutes of receiv- Route 199 north to Grants Pass and ing notice, he said, and those already a backcountry route up Deer Creek On July 30, Mike Lohrey’s type 1 packed should seriously think about on forest roads to Williams, OR. team moved into position near leaving immediately. The Illinois Valley Fire District Selma. A thick haze of smoke designated two fireproof schools hugged the ground, hiding the Gilpin took his seat knowing that standing in acres of cleared land. flames. The choking smoke sent even some of his colleagues found Firefighters and residents could take more Illinois Valley residents in his assessments alarmist. Lose shelter there if Florence crossed the search of clearer air. Motels in the whole Illinois Valley? See the highway. Grants Pass offering steep discounts to evacuees soon filled up. Florence Fire grow to 500,000 acres? It seemed implausible. Firefighters fanned out across the valley to protect 3,429 homes; 250 Local radio stations broadcast that But Gilpin knew that Florence had commercial buildings; and 2,200 night’s public meeting, but hundreds already forced fire managers to barns, sheds, and garages, assess- of people still gathered at the high throw out their normal strategies. ing each for defensibility against school to hear the latest fire news. Maps hanging outside told part of They’d long ago given up on the idea

Volume 65 • No. 2 • Spring 2005 11 manmade or natural. This was an audacious fire, he felt, one that could be vanquished only with audacious tactics.

Florence’s exploding growth was due in part to fire managers’ initial decisions to fight blazes elsewhere in the West that more directly threat- ened buildings and people. But now, Florence was a powerful monster, and the Nation’s best firefighters were here to stop it.

Lohrey had studied the map, memo- rizing the twists and turns in the terrain. He knew that if the fire wasn’t stopped immediately, it could Rogue River Hotshots fortifying a fireline for a burnout operation on the Biscuit Fire. become the most destructive wildfire Photo: Tom Iraci, USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002. in modern history, burning to the Pacific Coast, the California red- of controlling the boundaries of the One veteran wildland firefighter woods, and the Rogue River, threat- fire, and they had allowed Florence sitting nearby had no quarrel with ening the lives and homes of 50,000 to grow unchallenged in any direc- Gilpin’s grim assessment. Lohrey people. tion except east into the Illinois was the man chosen to rescue the Valley. They’d given up even trying threatened valley. And the next day Lohrey reviewed the previous plans to douse the head of the blaze. The at 6 a.m., he took command. for corralling Florence. They were goal now was simple: Save the com- textbook examples of how fires are munities of the Illinois Valley from a Shift in Strategy fought every summer in the United wall of flames, the advance of which At 5 a.m. on July 31, Lohrey crawled States, frontal attacks in which seemed inevitable. out of his two-man tent and took crews closed in on the flames and command of the fight against the created a fireline by clearing brush For years, Gilpin had preached Florence Fire. Immediately, he faced and chopping down trees. that losing a subdivision to wildfire a crucial question: Was sacrificing wasn’t a question of if, just a ques- another 547 square miles (1,417 The approach was sound, but Lohrey tion of when. But even he had never km2) of forest his best shot at stop- believed it would not work on this imagined losing the entire Illinois ping the blaze? blaze. The planned firelines had Valley. Never until July 30. been dug too close to Florence, a One of the Nation’s most experi- fire so powerful it could throw flam- In the previous 24 hours, the enced fire managers, Lohrey viewed ing missiles 2 miles (3.2 km) ahead. Florence Fire had gobbled another firefighting as a battle of wits. Every He began drawing a new map in 65,000 acres (26,000 ha), an expanse blaze had a personality, a charac- his head, one in which the fire, or two-thirds the size of Portland. It ter whose vulnerabilities could be the firefighters, would be allowed had run 5-1/2 miles (8.9 km) in just exploited. to burn through an unprecedented 90 minutes. At that rate, the fire amount of timber. Victory through could be in Selma in less than an But Florence was especially savvy retreat. hour. And as Gilpin sat in the dark- and strong. In nearly three decades ened gym, he figured there was a in the woods, Lohrey had seldom Lohrey believed in surrounding him- good chance that the Florence Fire seen a fire that combined its brute self with take-charge leaders who would enter the Illinois Valley in sev- force, unpredictability, and pre- would solve day-to-day problems on eral places and at a dead run. ternatural ability to leap barriers, their own. His job was to craft an

Fire Management Today 12 was not much for inspirational Along a 20-mile stretch of U.S. Route 199, the speeches, preferring to lead by exam- view to the west was an ominous string of plumes ple. This was his first day as the new towering over the valley. commander, a risky moment. Lives could be endangered by a missed communication. overall strategy and to set the tone, big map tacked to a plywood stand. even when events turned chaotic. Firefighters were preparing for a “Today the emphasis needs to be on “You have to be calm,” he said. “You massive burnout on a curved line making sure you give clear orders,” have to be the calm in the eye of the that stretched 30 miles (48 km) he told his aides. “And that you storm.” along the eastern edge of the fire. understand the instructions you’re The plan was to stop Florence’s given. It’s going to be a long season,” The First Day advance to the east by blackening he added. “It already has been.” nearly 200 square miles (518 km2) of Just before 6 a.m., Lohrey walked The First Burnout through the hodgepodge of tents forest. and equipment that the firefighters After the meeting, Lohrey and his had instantly dubbed Yurt City. At Lohrey had never come close to operations and planning chiefs the other end of the fire camp, 100 attempting a burnout of this size. gathered at the large map. He lis- firefighters in Nomex—branch direc- Once, in New Mexico, he had con- tened intently as they pointed to the tors, division superintendents, and sidered torching 39 square miles immediate dangers, but then his 2 crew bosses—gathered under some (101 km ), but the sheer scope of the eyes slid toward areas miles away tall ponderosa pines. They would, operation had given him pause. Then from the front line. It was Lohrey’s they hoped, lead Lohrey’s 1,546 fire- the winds had changed, and he had job to think days, even weeks ahead, fighters to the rescue of the Illinois chosen direct attack. to figure out every move the fire Valley. might make and to get there first. Everyone present understood the One by one, Lohrey’s chiefs delivered risks. A sudden shift in wind, and the Already, he’d sent a team of fire the latest intelligence on Florence’s burnout could turn on the firefight- managers to Gold Beach to prepare movements. The first report was bad: ers and the Illinois Valley. Lohrey a new fire camp for another type 1 A shift in the winds had postponed a burnout needed to protect Selma from the advancing flames.

On the plywood podium, a meteo- rologist recited the brutal weather forecast: sunny, hot, and a red-flag warning for northeast winds.

Then, Erik Christiansen, the team’s fire behavior analyst, went through the basic variables: wind, humid- ity, temperature, and the fire’s heat. All were at historic levels for the Siskiyou National Forest—condi- tions perfect for explosive fire spread. It had been several days since Florence had spun up a plume, but Christiansen said it could happen again by afternoon.

Lohrey’s operations chief described Firefighter firing off an incendiary flare to aid in a burnout operation on the Biscuit Fire. the day’s work ahead, pointing to a Photo: Tom Iraci, USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002.

Volume 65 • No. 2 • Spring 2005 13 team that would eventually protect communities on the Oregon coast. A sudden shift in wind, and the burnout could turn At dusk, his crews began burning the on the firefighters and the Illinois Valley. fireline that would protect Selma. The aim was to create a blackened band that would halt Florence in its the burnouts much tougher. Lohrey tracks. was racing the weather, as well as The line was huge—405 miles (652 Florence. km) long—and the area within it The wind fanned the flames on beyond imagining—500,000 acres the ground, pushing fire upslope The Sacrifice Line (200,000 ha), or 781 square miles 2 through grasses on the forest floor Tom Link was the acting district (2,023 km ), an expanse larger than and into brush. Here and there, a ranger on the Illinois Valley Ranger Washington County, OR. tree caught fire, flaming up with District when Florence reared its a brilliant flare that quickly disap- head. A career timber manager, Link Florence had already burned peared. Soon all the ridges northeast had hiked most trails in the old- through 165,000 acres (67,000 ha). of Selma glittered orange as stars growth forest. He knew its rivers, Lohrey was recommending that fire- sparkled in the night sky. streams, and peaks the way some fighters deliberately double or triple people know their home town. the amount of burned forest. Link The next morning, Lohrey’s crews looked glum. He pressed Lohrey to moved onto the blackened ground On August 2, Lohrey came to his save more. to make sure wind couldn’t whip office carrying a plan for letting the fire back to life. Timber cutters much of the surrounding forest Lohrey wouldn’t budge. The amount downed fire-weakened trees and burn. The two men spread a map of of territory to be surrendered had to snags that could fall on firefighters the national forest, and Link peered be huge because firefighters would working below. Ground crews fol- at the jagged circle drawn around need that much time to build a wide- lowed with water to douse logs still the edges that marked how much of enough cordon around the fire. Both burning on the ground and tree his district would have to be surren- men knew that the previous plans roots burning underground. Some of dered to quell this monstrous blaze. had drawn much tighter lines, which the firefighters removed their thick Florence had skipped over with ease. leather gloves and ran their hands along the ground, seeking tactile confirmation that the ashes were cold. When they were done, they had created a dead zone half a mile (0.8 km) wide, broad enough to stop Florence at its most fierce.

Meanwhile, Florence probed an entirely different corner of the map. The flames had licked to within 5 miles (8 km) of Agness, a hamlet near the confluence of the Illinois and Rogue Rivers.

At that morning’s briefing, the meteorologist forecasted cooler tem- peratures with a chance of rain. The report was not welcome. A splash of rain would not stop Florence; the fire was hot enough to dry its own fuel as it pushed through the Explosive charge detonating to clear a quarter-mile fireline on the Biscuit Fire. Photo: woods. Rain, however, would make Tom Iraci, USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002.

Fire Management Today 14 “Worst case is worst case,” he told Lohrey respectfully declined. ried Lohrey failed to materialize. The Link. “That’s what you have to think Shifting the burn line would leave temperature dropped, delaying some of.” open a backdoor route for Florence of the burnouts. to escape and encircle the Illinois “We’ve gone through a couple of Valley. On August 7, the firefighters marked worst cases already,” replied Link, a historic occasion. The bulldozers weary after 3 exhausting weeks fight- “I don’t want to leave the back door scraping the fireline northward from ing Florence. “We’ve had a mind- open,” Link agreed. the smaller Sour Biscuit blaze kissed expansion process in what we think blades with the bulldozers plowing is bad.” The situation was little better in south from the Florence Fire. Within the north. If the line was closer days, the fires united, merging into Lohrey needed the district ranger and Florence jumped it, the road to a single titanic conflagration. Lohrey to approve his plan, but Link was Grants Pass would be wide open. The and the fire managers retired the unpersuaded. Wasn’t Lohrey giving terrain would drive the fire toward name Sour Biscuit. At that point, Florence too much room? the town. the historic 2002 wildfires in south- western Oregon all became known as “The reality is, 500,000 acres Link traced the topographical map Florence. [200,000 ha] is not unreasonable with his fingers. Lohrey had a point. given the time of the year, and where “The ridges are all the wrong way,” The fire mounted one last charge we are in the season, and how much Link said. along the eastern front, just where fire we’ve got,” Lohrey replied. Lohrey had feared it might. The The debate was over. The sacrifice flames advanced toward U.S. Route “Fires are not unlike floods. People’s line had been drawn. 199, angling for a 2-mile (3.2-km) memories are short,” said Lohrey, hole in the line. This was the “back who compared Florence to a hun- door” that Link and Lohrey had wor- dred-year flood. Living memory—50 The amount of territory ried about in their conversation days years—“is a short timespan for the to be surrendered had earlier. way these things happen,” he said. to be huge because The following afternoon, Florence Link stared at the map silently, his firefighters would need charged the line. Fanned by north- eyes darting from ridge to peak to that much time to build east winds, she crossed the fireline creek to forest road. He nodded. a wide-enough cordon along a narrow 6-mile (9.6-km) Lohrey had his approval. stretch. Lohrey’s firefighters had around the fire. fire in their faces. For the first time That night, Link had second since Lohrey took charge 9 days ear- thoughts. He approached Lohrey lier, his crews retreated and began a after an evening meeting at Yurt City Slamming the new fireline. and walked him to the map. Would Back Door it be possible to slide the line in two Over the next 4 days, Lohrey’s crews Lohrey considered a daring counter- environmentally sensitive areas: the worked to build the first 30 miles (48 move: going on the attack against southeastern corner of the forest, km) of fireline. It was a staggering the implacable blaze. He now had a habitat for rare plants that grow project. No Federal fire manager had hotshots galore. But even the only in the Siskiyou Mountains; ever attempted to create a cordon of Nation’s best wildland firefighters and a stretch of land north of the this size. couldn’t attack a fire kicked up by Kalmiopsis Wilderness that was those northeast winds. home to 49 pairs of northern spotted Day and night, the burnouts con- owls? tinued, looping around threatened Two days later, the winds shifted homes near Selma, Kerby, Cave enough to allow a direct assault on “That’s a big burnout,” Link said, Junction, and O’Brien. Florence con- the fire. Helicopters dropped bucket referring to the plans for the south. tinued to test the defenses, throw- after bucket of water on the flames, “It’s a question I’m going to get from ing up columns that threatened to and then 80 hotshots moved off the the public. Why did we need to burn become full-fledged plumes. But forest road into the rocky drainage of out 8 miles [13 km]?” none did. The rains that had wor- Whiskey Creek, backed by four expe-

Volume 65 • No. 2 • Spring 2005 15 rienced Forest Service hand crews. The hotshots worked a 12-hour shift through the night, tearing down trees and brush with hand tools and chain saws. By daybreak, the line extended 2,000 feet (610 m).

Lohrey sent in fresh crews the next morning. The winds cooperated, and within 24 hours the back door had been slammed shut, completing the 30-mile (48-km) line.

The Illinois Valley was safe. For the first time since his arrival, Lohrey felt confident that he had the upper hand on Florence.

Lohrey left the valley on August 13 Sign posted along the road to fire camp on the Biscuit Fire, August 2002. Photo: Tom Iraci, and headed home to Portland. The USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002. fire was not yet fully corralled; much work remained to be done along its trees left standing in this blackened The problem has been compounded northern and western flanks. But the moonscape are all dead. Another 41 by 80 years of suppressing wildland most serious danger had passed. percent of the land was less severely fires, which have left many forests damaged, and ecologists think the choked with fuel. A wave of lightning Aftermath thinning will eventually improve for- strikes on a hot summer day can Fire managers declared the fire est health. About 20 percent of the overwhelm the fire crews assigned contained on September 5—54 land within Lohrey’s sacrifice line to snuff out small fires before they days after lightning had struck near did not burn at all. become conflagrations. Florence Creek. The official fire size, determined by aerial mapping, was Reevaluating Strategies set at 499,570 acres (202,169 ha). The experience of battling the An estimated 38 percent The worst-case scenario Lohrey out- Florence Fire calls into question of the 500,000 acres lined for Link in August had turned the priorities that drive national within the fire perimeter out to be all too true. firefighting. Longstanding policy requires protecting people and build- was reduced to charcoal In the end, the damage to property ings first. When Florence was burn- and ash. and people was startlingly small. ing in uninhabited timber, fire man- Florence claimed four summer cab- agers attacked other fires that posed ins, nine outbuildings, and the his- a more immediate threat to homes Fire managers have acknowledged toric Snow Camp Mountain lookout. and subdivisions. Usually, that sort that they were stretched thin by an of policy makes perfect sense. But in outbreak of wildfires across 11 States Experts are continuing to study this case, it allowed Florence to grow and missed an early chance to put Florence’s toll on the Siskiyou so big and hot that it became almost out the Florence blaze. Helitack National Forest. An estimated impossible to stop, eventually threat- crews and smokejumpers in the 38 percent of the 500,000 acres ening the entire Illinois Valley. region had already been sent to fires (200,000) within the fireline was elsewhere in the West. Fire manag- reduced to charcoal and ash. The

Fire Management Today 16 ers say that they are reevaluating ever been attempted anywhere in the His contest with Florence had been their strategies for attacking wild- United States and in circumstances closely fought, and Lohrey knew fires in light of what they learned that left no margin for error. that good fortune had played a during 2002 in Oregon. role. At the most critical moments, “There was at least a week or two the winds blew in exactly the right Today’s tinder-packed forests burn there,” Gilpin said, “that I felt there directions. “We could have been in hotter and faster, making traditional was a good chance we could lose trouble,” he said. “It could have been frontal attacks on fires more dif- either the Illinois Valley or a big por- a lot bigger.” ficult. The tactic that eventually tion of it.” contained the blaze in southwestern Lohrey feels certain the forests of Oregon—penning the fire within a A Flammable World the West will spawn more Florences. vast perimeter—involved surrender- Lohrey gave up more acres than The woods are packed with fuel, the ing more forest to the flames than firefighters were accustomed to sur- legacy of generations of suppressing had ever been accepted before. rendering, but he had no second fire. thoughts. Relinquishing a half-mil- The firefighters relied on a proven lion acres, he thought, was better for “It’s a flammable world,” he said. approach—burning the fuel in front the forest, cheaper for the taxpayers, “Fires need to burn and will burn.” ■ of the advancing flames. But they and safer for the firefighters. did so on a larger scale than had

WEBSITES ON FIRE* Partners in management, safety, education, risk of fire in their community. Protection and prevention; and by encourag- The site also provides two practi- Partners in Protection ing community-based initiatives. cal assessments to help individu- The threat of fire in the wild- Based in Alberta, Canada, Partners als and agencies determine the land/urban interface (WUI) is in Protection is a multidisciplinary fire hazard in a given part of the constant and real. Partners in group of professionals from gov- WUI. The new FireSmart multi- Protection helps those living in ernment associations and agencies media CD, which can be ordered the WUI by providing informa- responsible for emergency services online, includes a manual on tion about disaster services, fire as well as for land management hazard assessments and the and planning. video Wildfire! Preventing Home * Occasionally, Fire Management Today briefly Ignitions. The site also has use- describes Websites brought to our attention by the wildland fire community. Readers should not con- The Website provides tools to help ful links to agencies and organi- strue the description of these sites as in any way communities learn the principles zations involved in disaster relief exhaustive or as an official endorsement by the USDA Forest Service. To have a Website described, contact of FireSmart. Visitors can down- and fire education and safety. the managing editor, Hutch Brown, at USDA Forest load an interactive planning man- Service, Office of the Chief, Yates Building, 4th Floor Northwest, 201 14th Street, SW, Washington, ual—FireSmart: Protecting Your Found at learn about how to mitigate the

Volume 65 • No. 2 • Spring 2005 17 HOW DID PREFIRE TREATMENTS AFFECT THE BISCUIT FIRE? Crystal Raymond and David L. Peterson

ost scientific literature supports forest thinning Mto reduce the severity of wildland fires, but the effective- ness of thinning in modifying fire behavior has not been well documented. The Biscuit Fire of 2002 offered a great opportunity to study the effects of mechanical thinning on fire behavior during a megafire. The Thinning Theory Forest thinning is done to prevent surface fires from transitioning to crown fires. Theoretically, reduc- ing canopy fuels and eliminating ladder fuels will decrease the prob- ability that a crown fire will initi- ate and spread (Cron 1969; Omi and Martinson 2002; Pollet and Omi 2002; Scott and Reinhardt 2001; Stephens 1998).

Most studies on fuel treatment efficacy focus on forests with low-severity, high-frequency fire regimes, such as ponderosa pine and Douglas-fir forests in the Interior West. As a result of fire exclusion, these forests are dense, producing greater canopy and lad- der fuel loadings (Agee 1993). Aerial view of fire damage on thinned and untreated plots 1 month after the Biscuit Fire. At the opposite end of the spec- Thinned plots are circled in orange and untreated plots are circled in red. Photo: USDA trum are forests with high-sever- Forest Service, 2002. ity, low-frequency fire regimes, because weather influences fire Little information exists for forests such as Pacific coastal, subalpine, behavior more than fuels in these with mixed-severity fire regimes, and boreal forests. However, ecosystems (Bessie and Johnson which encompass a wide range of 1995; Turner and others 1994), fire frequencies, extents, and severi- Crystal Raymond is a graduate student in the Fire and Mountain Ecology Laboratory, thinning is less effective in reduc- ties. The result of this variability College of Forest Resources, University ing fire severity than it is in for- makes it more difficult to quantify of Washington, Seattle, WA; and David est ecosystems with a higher fire the ecological role of fire than in for- Peterson is a research biologist for the USDA Forest Service, Pacific Wildland Fire return frequency. ests with low-severity fire regimes. Sciences Laboratory, Pacific Northwest Research Station, Seattle, WA.

Fire Management Today 18 Biscuit Fire: A Research Opportunity In 1992, scientists from the USDA Forest Service’s Pacific Northwest Research Station started the Long- Term Ecosystem Productivity (LTEP) study on the Siskiyou National Forest in southwestern Oregon. The initial goal was to assess the effects of plant community composition and large woody debris on the processes that affect forest ecosystem produc- tivity. When the Biscuit Fire roared through the area on August 16, 2002, the goal changed.

The Biscuit Fire was Oregon’s largest recorded fire and one of the largest ever to occur on national Crown scorch of overstory trees on a thinned plot (left) and an untreated plot (right). The thinned plots studied suffered more damage in the Biscuit Fire than the untreated plots, forest land. It burned more than probably because there were more fine woody debris and dense hardwood sprouts, fuel- 499,000 acres (202,000 ha), and cost ing a more intense surface fire. Photo: Crystal Raymond, Fire and Mountain Ecology Lab, more than $150 million to suppress. University of Washington, Seattle, WA, 2003. Although weather contributed to fire severity through high temperatures, sified as a mixed-severity fire regime thinned plots. On eight treatment low nighttime relative humidity, and with a fire return interval of 90 to plots, tree crowns were removed dry east winds, the LTEP sites burned 150 years. together with the last log; on two under more moderate weather condi- others, crowns were left onsite. In tions. By the time of the Biscuit Fire, the fall of 2001, Siskiyou National the LTEP sites had been carefully Forest resource managers conducted studied, with plenty of prefire data prescribed burns in the understory collected. A wildland fire burning of thinned plots where crowns were Data are sparse for through the area was a great oppor- left onsite. forests with mixed- tunity to study the effects of thin- severity fire regimes ning on fire severity in forests with The Biscuit Fire burned through that encompass a wide mixed-severity fire regimes. three thinned plots, one thinned– range of fire frequencies, underburned plot, and two untreated Experimental plots with minimal torching of over- extents, and severities. Treatments story trees. Burned plots are from In the winter of 1996, 10 of the 27 2,690 to 3,610 feet (820–1,100 m) in elevation, with southeast and north– The LTEP sites were on the western LTEP treatment plots, each from 15 to 20 acres (6–8 ha) in size, were northeast aspects and slopes of from perimeter of the Biscuit Fire, about 15 to 40 percent. 12.5 miles (20 km) inland from mechanically thinned using a heli- copter to remove the logs. The other the coast. Douglas-fir, with a small Study Tactics amount of sugar pine and knobcone 17 plots were either clearcut or left Before the Biscuit Fire, researchers pine in the overstory, dominates the undisturbed to serve as a control. collected extensive fuels and vegeta- area. The subcanopy is composed of Thinning from below removed most tion data before and after thinning, hardwoods (tanoak, Pacific madrone, of the subcanopy hardwoods and including data on forest structure and chinquapin) and Douglas-fir. conifers, reducing tree density from and dead and down woody debris. These stands, established approxi- approximately 419 trees per acre Following the harvest, researchers mately 100 years earlier following a (1,035 trees per ha) to 85 trees per established five permanently marked stand-replacing fire, have seen little acre (210 trees per ha). Logging and mapped 0.08-acre (0.03-ha) tree active management. The area is clas- slash was minimally treated on the

Volume 65 • No. 2 • Spring 2005 19 plots per treatment plot. We tagged How Did Thinning burned plot, unable to spread all live trees and snags and measured Pan Out? through the sparse surface fuels. On diameter, species, crown class, tree the thinned plots, the fire consumed Our study encompassed only a few height, and canopy base height. the subcanopy layer of 5-year-old stands, and the sample size is not Stem mapping of trees in the plots hardwood sprouts and the extensive large enough for rigorous statistical helped us locate all trees following fine wood that resulted from thin- inferences about differences in treat- the fire. ning. Crowns of overstory trees were ments. Therefore, the data presented nearly 100 percent scorched. here should be considered limited In the summer of 2003, 1 year after and observational rather than statis- the Biscuit Fire, we again measured On the untreated plots, there was tically based. stand structure and fuels and collect- no evidence that subcanopy hard- ed additional data to assess fire dam- wood and conifer trees served as age to overstory trees. Tree damage ladder fuels. The older, larger hard- measurements included maximum A wildland fire burning wood trees were not consumed. The bole char height, maximum crown through a long-term untreated plots and the thinned– scorch volume height, crown scorch experiment with pre-fire underburned plots had much lower volume (in percent—a visual esti- data created a great quantities of fine woody debris prior mate), and percent cambium kill. We to the fire and lower consumption of extracted four cores per tree (uphill, opportunity to study the fuels during the fire. downhill, and two cross-slope sam- effects of thinning on ples) at 1.6 feet (0.5 m) above the wildland fire severity Lessons Learned ground to assess cambium status. Greater fire damage to trees from We tested each cambium sample radiant and convective heat rather for the presence of peroxidase, an Maximum bole char height and than crowning occurred in the enzyme found in all living plant tis- maximum crown scorch height thinned plots. Using Rothermel’s sue. One dead sample equates to were similar on all treatment plots. (1983) fire prediction models in approximately 25 percent cambium However, there were definite varia- hypothetical stands, Graham and girdling, two samples to 50 percent tions in crown scorch volume (fig. 1) others (1999) provided a theoretical girdling, and so forth. and some variations in percent cam- basis for greater surface fire intensity bium girdled (fig. 2). On the thinned resulting from residual slash fuels Trees often take several years to plots, overstory trees with a diam- and higher windspeeds in thinned succumb to fire damage, so mortal- eter at breast height of greater than stands. The high level of crown ity data collected a year after a fire 10 inches (24 cm) had the highest scorch within the thinned LTEP do not reflect total tree mortal- crown scorch volume and cambium plots most likely resulted from con- ity. However, previous studies of death. On the thinned–underburned vective heat rising from the intense fire-caused Douglas-fir mortality plots, overstory trees had the least surface fires below. The intensity of show that percent crown scorch crown scorch volume and cambium these surface fires was exacerbated and percent cambium kill are the death. On the untreated plots, over- in the thinned plots where there was most important damage variables story trees had moderate damage more fuel in the form of fine woody for predicting mortality (Peterson and the most variability in crown debris and dense hardwood sprouts. and Arbaugh 1988). These variables scorch volume and cambium death. These fuels were not present in the therefore allowed us to predict total untreated plots, and they were con- fire-caused tree mortality. Crown The Biscuit Fire burned through sumed in the prescribed burning scorch height and bole char height the thinned and untreated plots as treatment in the thinned–under- are more superficial damages and an intense surface fire but stopped burned plot. indicate less about long-term at the edge of the thinned–under- fire effects.

Fire Management Today 20 In forests with mixed-severity fire 0 20 40 60 80 100 regimes (especially where hardwoods Untreated Thinned/underburned Thinned are present), removing ladder fuels 100 might actually increase damage to the remaining stand in a subsequent 80 wildland fire. In the Biscuit Fire, the hardwood subcanopy affected fire behavior in ways other than serving 60

of total as a ladder fuel. It is possible that shading by the hardwoods slowed 40 the desiccation of dead surface fuels prior to the fire. During the fire, Percent 20 the hardwoods possibly decreased windspeeds in the untreated stands. In the thinned plots, consumption of 0 hardwood sprouts probably contrib- 0 20 40 60 80 100 0 20 40 60 80 100 uted to crown scorch, whereas the Percent crown scorch older hardwoods in the untreated plots might have prevented the Figure 1—Percent crown scorch in overstory trees on untreated, thinned–underburned, and upward movement of heat to over- thinned research plots following the Biscuit Fire of 2002. Damage was high on thinned plots, story conifers. mixed on untreated plots, and low on thinned–underburned plots. Observations suggest a two-step pro- 0 25 50 75 100 cess to prevent wildfires from crown- Untreated Thinned/underburned Thinned ing in forests with mixed-severity 80 fire regimes:

1. Thin dense stands to decrease 60 ladder fuels; and 2. Remove post-thinning slash and other accumulated surface fuels to of total 40 confine subsequent fire behavior to a relatively cool surface fire.

Percent 20 Although this two-step process is more time consuming and costly than thinning or prescribed burning 0 alone, it appears to be more effective in enhancing suppression efforts and 0 25 50 75 100 0 25 50 75 100 in reducing undesirable damage to overstory trees. Percent cambium girdled

Figure 2—Percent cambium girdled in overstory trees on untreated, thinned–underburned, and thinned research plots following the Biscuit Fire of 2002. Damage was higher on thinned and untreated plots than on thinned–underburned plots.

Volume 65 • No. 2 • Spring 2005 21 similar stand treatments on fire behav- ior in Western forests. Gen. Tech. Rep. PNW–GTR–463. Portland, OR: USDA Forest Service, Pacific Northwest Research Station. Omi, P.N.; Martinson, E.J. 2002. Effects of fuel treatment on wildfire severity. Final report to the Joint Fire Science Program. Boise, ID. Peterson, D.L.; Arbaugh, M.J. 1988. Estimating postfire survival of Douglas-fir in the Cascade Range. Canadian Journal of Forest Research. 19: 530–533. Pollet, J.; Omi, P.N. 2002. Effects of thin- ning and prescribed burning on crown fire severity in ponderosa pine forests. International Journal of Wildland Fire. 11: 1–10. Rothermel, R.C. 1983. How to predict the spread and intensity of forest and range fires. Res. Pap. INT–143. Ogden, UT: USDA Forest Service, Intermountain Forest and Range Experiment Station. Scott, J.H.; Reinhardt, E.D. 2001. Assessing crown fire potential by linking models of surface and crown fire behavior. Res. Pap. RMRS–RP–29. Ogden, UT: USDA Forest Service, Rocky Mountain Research Station. Stephens, S.L. 1998. Evaluation of the effects of silvicultural and fuels treatments on potential fire behavior in Sierra Nevada mixed-conifer forests. Forest Ecology and Management. 105: 21–35. Turner, M.G.; Romme, W.H.; Gardner, R.H. 1994. Landscape disturbance models and long-term dynamics of natural areas. Natural Areas Journal. 14: 3–11. ■ Fine fuel accumulation before (above) and after (below) the Biscuit Fire. Fine fuels accumu- lated following a thinning treatment after tree crowns were removed, but the Biscuit Fire almost completely consumed them. Photo: USDA Forest Service, 2003.

Acknowledgments References The authors thank Bernard Agee, J.K. 1993. Fire ecology of Pacific Northwest forests. Covelo, CA: Island Press. Bormann, Robyn Darbyshire, and Bessie W.D.; Johnson, E.A. 1995. The relative Colin Edgar for their assistance in importance of fuels and weather on the fire all aspects of the study. Research was behavior in subalpine forests. Ecology. 76: 747–762. supported by the Forest Service’s Cron, R.H. 1969. Thinning as an aid to fire Pacific Northwest Research Station, control. Fire Control Notes. 30: 1. Pacific Wildland Fire Sciences Graham, R.T.; Harvey, A.E.; Jain, T.B.; Tonn, Laboratory, Seattle, WA. J.R. 1999. The effects of thinning and

Fire Management Today 22 SNOW CAMP LOOKOUT: REMEMBERING A BISCUIT FIRE CASUALTY

Jan Robbins

onely souls, perched high in their lookouts, were once the For a lookout, the ever-changing view can bring Lfirst to spot forest fires and put raptors riding on afternoon thermals, Bolander’s lily the suppression machine into action. in bloom, or anvil clouds foretelling thunderstorms. Often a romantic stereotype, look- outs have become synonymous with the USDA Forest Service. Lookout History In 1958, the Forest Service replaced Snow Camp Lookout on the Siskiyou the original building with a 15-by- Forest fire lookouts have operated National Forest in Oregon was first 15-foot (4.6-m x 4.6-m) “modern” for more than a hundred years—the used during World War I, even lookout structure. The lookout was first was erected in 1902. Men, before a structure was built on the staffed through the 1972 fire season, women, and couples, some paid and site. The original building was con- then virtually abandoned for many some volunteers, took their turns at structed in 1924. During World War years. In 1989, the agency began this solitary activity. Some found the II, lookouts occupied Snow Camp renovating the building and, in 1990, job trying, but others thrived. year-round and scanned the hori- opened it to the public as a rental zon for enemy warplanes as well as cabin (fig. 1). smoke. On August 13, 2002, despite the efforts of firefighters, the Biscuit Fire incinerated Snow Camp Lookout (fig. 2). All the equipment, maps, lumpy beds, outhouse, and everything that made Snow Camp Lookout a home to so many for so many years burned to the ground.

In some ways, the panorama from the top of Snow Camp Mountain was unchanged by the fire. Clefts formed by rivers and streams, the shape of the mountains and hills, and some of the vegetation remain untouched. And eventually, of course, the forest will again be whole.

Figure 1—Snow Camp Lookout after renovation in 1989. Photo: USDA Forest Service.

Jan Robbins is a hydrologist for the Siuslaw National Forest, Walport, OR.

Volume 65 • No. 2 • Spring 2005 23 Place of Beauty Snow Camp Lookout was a place of growth, beauty, and quiet joy. It provided a unique opportunity to enjoy the patterns of life and land- scapes while nurturing relationships in relative comfort and privacy. In the confined space of Snow Camp, friendships were deepened, anni- versaries celebrated, family ties strengthened, and endurance tested.

A fire can destroy many things, but memories are fireproof and can sometimes give new life to the past. Dedicated volunteers, some who remember and some yet to experi- ence, are rebuilding the Snow Camp Lookout. To the delight of many, the Forest Service will again offer the Figure 2—The remains of Snow Camp Lookout following the 2002 Biscuit Fire. building for public rental and enjoy- Photo: USDA Forest Service. ment. ■

CONTRIBUTORS WANTED We need your fire-related articles and photographs for Fire Management Today! Feature articles should be up to about 2,000 words in length. We also need short items of up to 200 words. Subjects of articles published in Fire Management Today include:

Aviation Fire science Preparedness Communication Fire use (including prescribed fire) Prevention/education Cooperation Fuels management Safety Ecosystem management Firefighting experiences Suppression Equipment/technology Incident management Training Fire behavior Information management (including sys- Weather Fire ecology tems) Wildland/urban interface Fire effects Personnel Fire history Planning (including budgeting) To help prepare your submission, see “Guidelines for Contributors” in this issue.

Fire Management Today 24 PREDICTION ERRORS IN WILDLAND FIRE SITUATION ANALYSES

Geoffrey H. Donovan and Peter Noordijk

ildfires consume budgets surable objectives are established. nomic, environmental, social, and and put the heat on fire Criteria include firefighter safety, other. For each alternative, users Wmanagers to justify and potential resource damage, and calculate an expected objective score control suppression costs. To deter- loss of private structures. by multiplying the probability of mine the appropriate suppression 2. Plausible suppression alterna- each outcome by its objective score, strategy, land managers must con- tives are developed. All alternatives then summing the resulting scores. duct a wildland fire situation analysis have a strategic plan focusing on The expected suppression cost and (WFSA) when: firefighter and public safety, avail- resource damage are calculated in able resources, containment time, the same way. Users display esti- • A wildland fire is expected to or probability of success, final fire mates in a decision tree format does escape initial attack, size, resource damages, and sup- (fig. 1). • A wildland fire managed for pression costs. resource benefits exceeds prescrip- 3. Suppression alternatives are ana- Predicted Versus Actual tion parameters, or lyzed. Managers assess the compat- Outcomes • A prescribed fire exceeds its pre- ibility of a suppression strategy We obtained data from the USDA’s scription and is declared a wildfire. with forest plan objectives, safety, National Information Technology and probability of success. Then Center on fire size and suppression On large wildfires, land managers they select the alternative that cost for 157 wildfires handled by type sometimes conduct five or more minimizes resource damages and 1 and type 2 incident management WFSAs. suppression costs. teams during the 2002 fire season. Since WFSA data are not collected Although the WFSA process is Managers use several variables and nationally, we contacted local land important and land managers are probability estimates to assign an managers responsible for each of the required to use it, research on the overall score to alternative suppres- 157 fires and requested their com- accuracy of WFSA predictions is sion strategies. The alternatives eval- pleted WFSAs. We received WFSAs lacking. We used data from the 2002 uated are typically associated with for 49 fires—42 fires with only one fire season to determine how WFSA- either the objective (minimizing fire WFSA and 7 with multiple WFSAs, predicted outcomes compared to size, suppression costs, and so on) for a total of 58 WFSAs. The 49 fires actual outcomes in terms of final fire or the suppression strategy (direct varied in size from 87 acres (35 ha) size and suppression costs. attack, indirect attack, and so on). to about 151,000 acres (61,000 ha). Most WFSAs evaluate two or three The fires burned a total of almost What Is a WFSA? different alternatives. For each alter- 805,000 acres (326,000 ha) at a sup- The WFSA process is not prescrip- native, users define a target and a pression cost of about $312 million, tive. Instead, it is a decision analysis worst-case outcome, and they might for an average per-acre suppression tool that requires land managers to also define an intermediate fallback cost of about $388. evaluate different suppression strat- outcome. For each outcome, users egies. There are three stages of a estimate the probability, associated To determine the accuracy of WFSA WFSA (MacGregor 2000): suppression costs and resource dam- probability estimates, we compared ages, final fire size, and objective the estimated probability of target 1. Criteria for evaluating suppression score. outcome success to the actual out- alternatives are identified and mea- come. We asked two questions: An objective score indicates how well a particular outcome meets a series 1. Was actual fire size smaller than Geoffrey Donovan is a research forester and of objectives. Although users can Peter Noordijk is a research assistant, USDA target size? Forest Service, Pacific Northwest Research define their own objective categories, 2. Was actual suppression cost less Station, Portland, OR. the default categories are safety, eco- than target cost?

Volume 65 • No. 2 • Spring 2005 25 Across all WFSAs, the mean estimat- fires, the mean estimated probability The differences between estimated ed probability of success of the target of the worst-case outcome occur- worst-case and actual per-acre costs outcome was 71 percent. Actual fire ring was 16 percent. Actual fire help explain why worst-case cost data showed that fire size was small- data showed that fire size exceeded probability estimates are approxi- er than target fire size for 63 percent worst-case fire size by 7 percent and mately correct. For fires smaller and actual costs were lower than that actual costs exceeded worst-case than 18,000 acres (7,300 ha)—39 of target costs for 44 percent of the costs by 19 percent. Fire managers the 49 fires in our sample—worst- WFSAs. Fire managers often under- tended to overestimate the prob- case per-acre costs underestimated estimated the probability of a given ability of actual fire size exceeding actual per-acre costs. The tendency wildfire exceeding its target size and worst-case size but to slightly under- of fire managers to overestimate costs, although they were better at estimate the probability of costs the probability of a fire exceeding predicting fire size than costs. exceeding worst-case costs. its worst-case size is equivalent to overestimating fire size. If an over- The seemingly inconsistent results estimated fire size is multiplied by We used data from the are perhaps attributable to how an underestimated per-acre cost, the 2002 fire season to fire managers estimate suppression two errors in estimation usually can- costs. We hypothesize that fire man- cel each other, making worst-case determine how WFSA- agers estimate suppression costs by cost estimates approximately correct. predicted outcomes first estimating fire size and then compared to actual estimating per-acre cost, with sup- This hypothesis also explains why outcomes. pression cost being a product of the target cost estimates are less accu- two. To determine the accuracy of rate than target size estimates. per-acre cost estimates, we compared Figure 2 shows that for all fires, We used a similar process to deter- actual, worst-case, and expected per- estimated target costs underestimate mine the accuracy of worst-case acre costs (fig. 2). actual per-acre costs. The tendency probability estimates. Across all of fire managers to underestimate the probability of a fire exceeding its target size is equivalent to under- estimating fire size. If an underes- timated fire size is multiplied by an underestimated per-acre cost estimate, the result is a cost estimate Target outcome that underestimates actual cost more 75% frequently than target size underesti- Size: 7,500 acres mates actual size. B.Indirect Contain: 7 days Control: 10 days Fallback outcome The data in figure 2 raise the ques- 85% Score: 2.2 tion of why there is a difference in Exp. score 1.7 Size: 10,000 acres Cost + Loss: $11.3M per-acre cost between target and Exp Cost+Loss $13.9M Contain: 9 days worst-case outcomes. Fire size, Control: 12 days 25% fuel type, topography, weather, and Fallback outcome Score: 0.0 Cost + Loss: $15.1M resource availability can affect per- 15% acre costs. However, why would Worst-case outcome labeling an outcome as target or Size: 25,000 acres worst-case affect per-acre costs? Score: 1.4 Such a classification is subjective Cost + Loss: $54.4M and does not affect fire behavior, val- ues at risk, or resource availability.

It appears that the optimism associ- Figure 1— This wildland fire situation analysis decision tree was used to evaluate an indi- rect suppression strategy for the Grizzly Complex Fire in Lake County, OR. Other alterna- ated with target outcome estimates tives (such as A. and C.) included direct attack. Percentages are probabilities of realizing a also affects per-acre cost estimates, particular outcome. which for the majority of fire sizes

Fire Management Today 26 example, Lichtenstein and others Fire managers often underestimated the (1978) asked people to estimate the probability of a given wildfire exceeding its frequency of various causes of death in the United States. Respondents target size and costs. consistently overestimated the prob- ability of dying from unlikely causes such as tornadoes or food poisoning while consistently underestimating 10,000 the probability of dying from likely causes such as heart disease or can- cer.

1,000 Past research also suggests some possible ways to improve estimates. Baron (2000) has shown that when people first list the factors that they 100 believe will influence the probability of an event occurring, their subse- Cost per acre quent probability estimates are more accurate. Perhaps fire managers 10 should first list factors that could influence the success of a particular 100 1,000 10,000 100,000 strategy—such as weather, resource Fire size in acres availability, and topography—before estimating the probability of the strategy succeeding. We also found Target outcome Actual outcome Worst case outcome that managers tend to estimate fire size more accurately than sup- Figure 2—Per-acre costs as a function of fire size for target outcome, worst-case outcome, pression costs. Using available his- and actual outcome. The trend lines do not imply a strict linear relationship between per- torical per-acre fire costs might help acre costs and fire size. The general relationship between fire size and per-acre cost is suf- ficient; therefore, individual data points were excluded for clarity. improve suppression cost estimates. References are lower than worst-case per-acre Improving Estimates Baron, J. 2000. Thinking and deciding. cost estimates. In addition, target Cambridge, UK: Cambridge University Estimating probabilities is chal- per-acre cost estimates decline Press. lenging. Our results are consistent Lichtenstein, S.; Slovic, P.; Fischhoff, B.; more with increasing fire size than with previous research showing that Layman, M.; Combs, B. 1978. Judged worst-case per-acre cost estimates. frequency of lethal events. Journal of people tend to underestimate the In contrast, worst-case per-acre cost Experimental Psychology. 4: 551–578. probability of likely events occur- estimates are higher and show little MacGregor, D.G. 2000. Accounting for wild- ring and overestimate the probability land fire costs in wildland fire situation decline with increasing fire size. of unlikely events occurring. For analysis. Eugene, OR: Decision Science Research Institute. ■

Volume 65 • No. 2 • Spring 2005 27 FIRE IN OUR MOUNTAINS— AND MOUNTAINS IN OUR RIVERS*

Thomas W. Swetnam

familiar statement in the large. Typically, these fires burned changes in the fuels. The frequent media these days goes some- through ponderosa pine forests low-intensity fires were eliminated A thing like this: “It’s the larg- about once or twice per decade. This between 1890 and 1900 in most est recorded fire in the history of high frequency of fire occurrence Arizona forests when large numbers the State.” There is both truth and ensured that the flame lengths were of sheep and cattle began to feed on historical naiveté in that statement. low—about 1 to 3 feet (0.3–0.9 m) the grass that was so important to Our shock over the size of these in height—because dead branches the spread of surface fires. Livestock conflagrations highlights a growing and other fuels were regularly con- numbers fell after World War I, but problem: We are faced with a serious sumed and so did not accumulate. then government agencies began to threat to our current course of popu- The large, mature pine trees were be more effective at detecting and lation growth and development in left unharmed, except for a few that putting out wildfires. the West, but we continue to ignore incurred “fire scars” at their base. key lessons of the past. The truth It is from tree-ring dating of these Fire Impacts Today is that part of what we are witness- fire scars that we find the same fire During the rest of the 20th century, ing is new to our written histories. dates repeated over and over from living and dead fuels accumulated to At the same time, this assertion one mountain range to another the point that surface fires now can of “largest in history” ignores the throughout the Southwest. Burning rapidly “ladder up” into the canopies essential fact that much larger areas unhindered for months, fires swept of large trees. The very high intensi- burned in the not-so-distant 19th over millions of acres in 1748, 1851, ty and extent of these crown fires are century. How quickly newspapers 1879, and 1882. The big fire years unprecedented in southwestern pon- forget their own words. On April 16, of the 20th century—1994, 1996, derosa pine forests. One of the most 1882, for example, the Arizona Daily 2000, and 2002—are impressive to worrisome indications is the erosion Star noted that, “Prairie and wood modern Arizonans, but in terms of and flooding that often follow these fires have been raging in southern area burned, they pale in comparison crown fires. Loss of ancient soils Arizona and western New Mexico to the big fire years of previous cen- and formation of deep arroyos have recently. The territory burned over is turies. occurred following recent fires in reported to cover forty miles square southern Arizona and elsewhere in [about 1 million acres or 400,000 the West. ha]....” The big fire years of the Now add two more factors to this Historical Fires 20th century pale in comparison to the big mix of dense forests choked with There are numerous other historical living and dead fuels: people and accounts by soldiers and pioneers fire years of previous drought. People love to live in the of gigantic burned areas in the centuries. midst of pine forests. The attraction West during the 19th century, and is akin to our desire to build along we know from tree-ring studies picturesque flood plains or hurri- that fires were frequent and very And yet, the big fires we are witness- cane-prone coastlines. People usually ing today are different in important say, “Yes, I know it’s a risky place, ways from those of the past. Contrast but it’s not that risky, and I’ll take Tom Swetnam is a Professor of the image of low-intensity flames a my chances.” But of course it’s not Dendrochronology at the University of few feet in height sweeping through Arizona in Tucson, AZ, and Director of the just their “chance,” and they won’t Laboratory of Tree-Ring Research at the grass and pine needles, with the bear the costs alone when the worst university. image we see today of an inferno scenario plays out. This kind of risk with flames 200 feet (60 m) high taking involves all of us, and it is * This article first appeared as an opinion piece in the rolling through the crowns of large Arizona Republic and the Arizona Daily Star on June 25, particularly dangerous in regions 2002. trees. The difference is caused by

Fire Management Today 28 such as the Southwest, where in dense forests, and our increasing There is a century of collective and drought is a fact of life. dependence on surface water from cumulative responsibility for this the Colorado River and Rio Grande. problem, and it is time to get over The year 2002 is panning out to be It’s not a pretty picture. the blame game. Land manage- one of the driest in a century, and if ment agencies can move beyond the we look to history we are reminded Doing Something “analysis paralysis” and litigation in that it could get worse. Both rain About It the courts by focusing their forest gauge and tree-ring records tell us All of this “sky-is-falling” fretting thinning work on the small-diameter that the worst droughts in the past does not mean that we can’t do any- trees. People living in the forest can 1,000 years have tended to come in thing about our predicament. We take responsibility for the risks they strings of dry years. A multiyear dry can save some of our forests by get- have exposed themselves and us to spell in the 1950s, for example, led ting on with the job of reducing the by removing fuels from around their to massive forest fires, bankruptcy of fuels. We should keep in mind that houses and by replacing flammable livestock ranches, and water ration- this problem does not extend to all roofs with metal. It is also time to ing in New Mexico and Arizona. forests, and small-diameter trees in enact policy, zoning, and insurance President Eisenhower declared New pine forests are the main type of tree measures in “fire plains,” just as we Mexico a disaster area and allocated that needs thinning. This is going do in flood plains. If we intend to Federal relief. Consider the con- to take all the tools we have at our keep catastrophic fire off our moun- sequences today of a 1950s-mag- disposal, including chain saws and tains—and our mountains out of our nitude drought in the Southwest, prescribed fire, and a massive effort rivers—we must find a way to live with greatly increased populations, (and many dollars) for many years to within a changing environment.■ numerous housing developments come.

Homes incinerated by the Rodeo–Chediski Fire. When it comes to policy, zoning, and insurance measures, it’s time to treat our western “fire plains” just like we do our flood plains. Photo: Tom Iraci, USDA Forest Service, Pacific Northwest Region, Portland, OR, 2002.

Volume 65 • No. 2 • Spring 2005 29 TREATMENT SUCCESS ON THE RODEO– CHEDISKI FIRE

Paul Keller

isitors to Arizona’s White Mountains come for the “Most places where we were able to thin and burn Vtrees—beautiful, fragrant in the last 10 years still have green trees.” ponderosa pines. Sadly, many trees –District Ranger Ed Collins now resemble charred toothpicks stuck into a scorched and barren landscape. They stand in testimony National Fire Plan will survive.” Shaking his head, he to Arizona’s largest recorded wildfire. Success added, “But a total of 32 percent of my ponderosa pine habitat was The Rodeo–Chediski Fire charged I took a drive with District Ranger completely burned in the fire. That’s across the land in the summer of Ed Collins into the burn’s interior what really hurts.” 2002, decimating life above and on on the Apache–Sigreaves National the ground. Forest. On Collins’s Lakeside Ranger National Fire Plan funding helped District, the Rodeo–Chediski Fire Collins implement the fuel-density- Thankfully, due to appropriate forest charred 48,300 acres (19,546 ha), or management activities that absorbed management, many trees did survive almost 20 percent of the district’s the fire’s wrath. “When a wildfire has (fig. 1). 270,000 acres (110,000 ha). that much energy, there’s nothing to actually stop it. But look how this “Most places where we were able to treated area here got the fire out of thin and burn in the last 10 years Paul Keller, a former hotshot and jour- the tops of the trees and back down still have green trees,” said Collins. nalist, is a contract writer/editor for the on the ground. By golly, it worked!” USDA Forest Service’s Fire and Aviation “Where we thinned and chipped, Collins motioned to a stand of pon- Management Staff, Washington Office, most of the trees also made it. Many Washington, DC. derosa pines where flames touched the trunks but left the tops resplen- dent with green needles. “See, it didn’t hurt these trees.” More Treatments Needed A few miles down another drought- stricken, dusty forest road, Collins pulled his rig over.

“Look here,” he said. “Over on this side of the road where we had already done thinning and prescribed fire, the trees still have their green needles—they’re going to make it. If we could only do more of this thinning on a landscape scale. Right now, we’re way behind the power curve. It’s really sad.”

Figure 1—Ponderosa pines in the foreground, after thinning and prescribed burning, survived the Rodeo–Chediski Fire. District Ranger Ed Collins surveys the fire’s devastation He pointed to the opposite side of in the background, where no treatments occurred. Photo: Tom Iraci, USDA Forest Service, the road, where treatments had not Pacific Northwest Region, Portland, OR, 2002 been implemented, and shook his

Fire Management Today 30 Not everyone was so lucky. Here and there, houses had completely disappeared—consumed by the fire. However, many homes survived thanks to fuel treatments around their perimeters (fig. 2). Beddow said that thinning and treatments in the surrounding forested areas also influenced the fire’s pattern.

Many homes survived the Rodeo–Chediski conflagration, thanks to treatments around their perimeters.

Figure 2—The Rodeo–Chediski Fire burned around this house but left it untouched, despite adjacent ornamental trees. The wide circular defensive space, coupled with adjacent cleared, “We know something about the open spaces, probably saved the home. Photo: Tom Iraci, USDA Forest Service, Pacific dynamics of fire behavior,” he said. Northwest Region, Portland, OR, 2002. “These treatments help restructure the fire’s burning characteristics. head. Thickets of tree boles, now and help stabilize the soils. So the They can manipulate the fuel–air charred the color of obsidian, leaned treatments still worked for us here, ratio in the canopy.” up from a parched, bare earth. too.” Where a forest once thrived, there Lesson Learned was now no life to be seen. Thankful Residents Less than a mile from where the fire When Tom Beddow, fire man- completely incinerated dozens of Of course, due to landforms and the agement officer on the Apache– homes, Beddow walked into a stand brute energy of the Rodeo–Chediski Sitgreaves National Forest, drove of ponderosa pine that had been Fire, in some areas the thinning his green Forest Service rig through thinned, logged, and understory- treatments could not save the trees. the small subdivisions sprinkled burned. Although the fire swept But the treated areas often knocked among Forest Service lands, the through this area, it never crowned the fire down enough to leave brown thankful residents waved and smiled. and the trees survived. needles on the trees—unlike the Others stopped their vehicles and completely denuded trees where the approached Beddow’s idling truck. The lesson was clear, for homes and fire was more severe. They all thanked him for protecting ponderosa pine ecosystems alike: their homes. “You guys did a tre- Treatments work, and we can’t afford “From a watershed point of view, mendous job,” said one man. “You to put them off. Surveying the dam- these brown needles are still good saved my home.” for the ecosystem,” Collins contin- age, Ed Collins summed it up. “This ued. “They will fall to the ground pains me,” he said. “So much of this damage could have been lessened.” ■

Volume 65 • No. 2 • Spring 2005 31 OPERATION SUCCESS: COLUMBIA SPACE SHUTTLE RECOVERY

Traci Weaver

hen the Space Shuttle Columbia exploded in the The space shuttle disaster was a national WTexas sky on February disaster, but the Columbia recovery mission was 1, 2003, it ignited emotions and widely seen as a success. burned a memory into the collective American mind. The space shuttle disaster initiated one of the largest Emergency Management Agency “Initially, some questioned why the search efforts in U.S. history, involv- (FEMA) made sense. Texas Forest Service was given a lead ing more than 25,000 people and role in Columbia recovery,” Stanford 450 organizations from all over the Texas Forest Service said. “The answer was simple. Most country. Largely due to an efficient Providing 20 incident manage- of the shuttle material was recov- Incident Command System (ICS), ment teams and nearly 600 hand ered on private lands. We have the the wildland fire community remem- crews, the Texas Forest Service soon authority to enter private property.” bers the 100-day Columbia Space became a key player in the shuttle Texas Forest Service liaisons worked Shuttle recovery effort as more than recovery effort. The search was the with each of the incident manage- a tragedy—it was also a successful first incident under the new U.S. ment teams to contact landowners, recovery mission. Department of Homeland Security gain access to private land, and col- and one of only a few Federal inci- laborate with local governments and Firefighters: dents in which a State agency had businesses. A Crucial Component a lead role. The goals of the Texas National Aeronautics and Space Forest Service during the recovery The Search Administration (NASA) officials, who mission were to ensure public safety, In 3 months, search crews located were initially skeptical about includ- recover the crew, retrieve evidence, 85,000 pounds (38,600 kg) of materi- ing firefighters, soon praised their and provide public assistance. al. That was nearly 38 percent of the efforts. reentry weight of Columbia, exceed-

“Using wildland firefighters with the incident management teams was the perfect tool for the job,” said Sam Ortega, NASA deputy lead onsite representative. “From the overhead to the individual crews, I never saw anything less than pure dedication to the mission. They are the reason that we are able to start a return to flight effort so quickly.”

“It became evident early on that it was going to require a long-term organized effort beyond what vol- unteers could sustain,” Stanford said, adding that recommending fire crews to NASA and the Federal

Searchers sweep through a field looking for debris following the Columbia Shuttle disaster Traci Weaver is a fire prevention specialist in February 2003. Photo: Jan Amen, Texas Forest Service, Lufkin, TX, 2003. for the Texas Forest Service, Granbury, TX.

Fire Management Today 32 ing the National Transportation Safety Board’s initial estimate that From the beginning, it was apparent that crews would recover less than 20 technology would play a vital role in tracking the percent. Searchers combed 700,000 shuttle material and mapping the searches. acres (280,000 ha) by ground and 1.6 million acres (540,000 ha) by air. In the East Texas counties of Search Challenges that link, it could have been easy for San Augustine and Sabine, workers An unusually wet Texas winter several entities to be operating in recovered the remains of the crew in prompted FEMA to approve extra their own bubbles.” just 12 days. boots and boot-drying trailers for searchers. Much of the search area FEMA and the U.S. Department Enough crucial pieces of the shuttle was thickly wooded and covered with of Homeland Security are moving were discovered for investigators to briars. As spring approached, snakes toward ICS, and the Environmental pinpoint the cause of the accident. and severe weather, including torna- Protection Agency has since adopted A Florida interagency hand crew in dos, joined the list of hazards. the system (see the sidebar on page Sabine County found the OEX data 34). recorder, or “black box,” which had “Despite adverse weather conditions, been missed during two earlier vol- our crews searched every day but ICS Flexibility unteer searches. two, and that was when Corsicana Flexibility during large-scale was hit with an ice storm,” Stanford efforts is crucial, according to Paul “Our crews managed to search said. “We were under time con- Hannemann, Texas Forest Service the entire targeted area along the straints, too. We tried to cover as chief regional fire coordinator. flight path—a 2-mile-wide [3.2- much ground as possible before km-wide] strip stretching from greenup, because we knew it would “All-risk incidents are not like fire,” west of Corsicana to Toledo Bend be more difficult to find material he said. “Every one of them is differ- Reservoir,” said State Incident once trees leafed out and everything ent. Teams must be flexible. ICS is Commander Mark Stanford, Chief began to grow.” simply a management tool. It’s our of Fire Operations for the Texas responsibility to stay focused on the Forest Service. “Searchers traversed Weather and critters were not the mission, be flexible, and be respon- a mind-boggling 552,000 miles only dangers. Each crew was briefed sive to changing demands.” [888,000 km].” on hazardous materials. Fifty poten- tially hazardous items were recov- Flexibility helped bring FEMA on Early search points were based on ered. Searchers were also told not to board, said Wayne Fairley, FEMA’s where other key elements of the key their radios within 50 feet (15 chief of operations for the Columbia shuttle were found. By March 1, m) of suspected explosive material. recovery. “We accepted a system that NASA had opted for a grid search No one was exposed or injured. was a little rigid to us,” Fairley said. along the flight path. Incident man- However, by being flexible, firefight- agement teams were tasked with Because many key agencies were ers allowed the system to accept planning the searches 24 to 48 hours unfamiliar with ICS, organizational FEMA procedures. in advance to allow for ample notifi- struggles developed. Those with a cation of landowners. basic understanding of the system “The ICS format that was used in were able to focus efforts more the shuttle recovery provided the “Parts of the incident were routine, quickly. working avenue to bring a lot of very like briefings and planning meet- different agencies together to accom- ings,” Stanford said. “Some things “Understanding ICS and how wild- plish one mission,” said Fairley, who were unique, like the search-and- land fire agencies operate helped has been involved in recovery efforts rescue training all of the crews had me early on,” said geographic infor- for 68 major national disasters. to undergo. Initially, NASA wanted mation system (GIS) coordinator “This incident brought out the most the search tight enough to produce Ronald Langhelm. “Based on our unique interagency cooperation I’ve a 90-percent probability of detection. experience in New York City after seen in my entire career.” They realized that was going to take 9/11, we immediately put together too long, so they relaxed it to 80 per- an interagency GIS team, which got Organization of the disaster field cent.” us started on the right foot. Without office (DFO) illustrates the impor-

Volume 65 • No. 2 • Spring 2005 33 Other Agencies work as one corporation with 10 but we absolutely believe it’s the Learn Value field offices.” right thing to do.” of Incident The EPA had been leaning toward The recovery effort also encour- Command ICS for large-scale incidents, using aged the Federal Emergency the USDA Forest Service playbook Management Agency (FEMA) to System for guidance. “We had learned incorporate ICS, although the about ICS from the wildland fire agency’s version of the system Three major national incidents service, but most of the exposure is tailored, said Wayne Fairley, prompted the Environmental we’d had was academic. We hadn’t FEMA’s chief of operations for the Protection Agency (EPA) to begin truly lived it,” Mullins said. “The Columbia recovery. adopting the Incident Command shuttle recovery propelled us a System (ICS), with Columbia recov- long way down the track to affirm- “Because our disaster responses ery serving as the final clincher. ing that ICS is what we wanted to are so very different—floods, tor- do. Living and seeing ICS in action nadoes, tropical storms and hur- “The anthrax scares, 9/11, and the was much more effective than ricanes, terrorism, etc.—FEMA had space shuttle disaster, all coming reading about it in a book.” to incorporate an ICS that was just within a 17-month period, con- a little more flexible than that used vinced the EPA that it had a lot to While the EPA has learned a lot by the Forest Service,” Fairley said. learn about crisis management and about ICS from wildland fire man- large-scale operations,” said EPA agers, growing pains still exist, Modified or not, Mullins is opti- Team Leader Jim Mullins. Mullins said. “Once there is an mistic that ICS can work. “It takes incident, and there’s an incident mountain-sized problems and “We had to learn to work well management team in place, you breaks them down into something together nationally, not just with work for the incident commander we can all get our arms around. other agencies, but also with our and not your normal boss, in an The Columbia recovery was a per- own regional offices,” Mullins said. uncertain situation, for an uncer- fect example of that.” “We used to function as 10 separate tain amount of time,” he said. “It’s regions. Now we know we need to been a cultural shift for the EPA, tance of flexibility. Originally, FEMA Service aviation specialist. Three set up DFOs in three locations. Role of Air Search others onboard the helicopter were Later, the three DFOs were consoli- The air search was another unique injured. dated into one central location in aspect of the mission. “NASA was Lufkin, TX. actively involved in the air opera- Sharing Lessons tions,” Stanford said. “The need for Learned “We were often asked why we didn’t an air search was apparent early “It’s important to carry forward and set up area command,” Stanford on because the ground crews were share the lessons learned,” Stanford said. “It was because this incident only searching a 2-mile-wide [3.2- said. “By doing our jobs more effi- was unique.” The DFO was a mul- km-wide] strip on either side of the ciently and effectively, we allow tiagency coordinating group, and flight path. The air search covered Americans to begin the healing pro- it was carrying out the functions of an additional 5 miles [8 km] on cess, just like the wildland fire agen- area command, he said. “But what either side of the flight path. Team cies’ efforts helped NASA in its quest was really unique was that it had an transitions would have interrupted to return to flight.” operational role of coordinating the air operations.” search.” The need to provide a continued However, the air search, which high level of national security will Astronauts at the DFO advised the involved 5,000 flight hours, came likely mean that the skills and GIS lab about search locations. The with a high price. A Bell 407 heli- resources exemplified by wildland GIS lab then developed the maps copter crashed on March 27, killing firefighters during the Columbia that were delivered to the base contract pilot Jules “Buzz” Mier recovery mission are destined to camps. and Charles Krenek, a Texas Forest become increasingly vital. ■

Fire Management Today 34 SHUTTLE RECOVERY: LARGEST GIS EMERGENCY RESPONSE TO DATE

Traci Weaver

any aspects of the Space Atwood’s first big challenge was try- Langhelm and John Perry, also from Shuttle Columbia recovery ing to organize the members of the FEMA, organized the GIS lab at the Mwere overwhelming from different agencies involved, who DFO in Lufkin by function—data the start—the size and scope of the were “all yelling loudly … that their management, product development, disaster, the number of agencies and project was top priority and not to and product function. A special team organizations involved, the urgency do anyone else’s,” he said. was brought in to handle technol- to find crewmembers, and the need ogy management so that servers, to determine the cause of the acci- Ron Langhelm from the Federal printers, and networking were all dent. Good management, technol- Emergency Management Agency properly functioning. Assignments ogy, and organization were essential. (FEMA) arrived later in the after- were made to cover printing, data noon to coordinate the interagency input, data presentation, and focus Gathering the Troops GIS. Having coordinated the GIS areas. The Environmental Protection Investigators had to recover as efforts for the 9/11 World Trade Agency (EPA) and NASA assigned much shuttle material as possible Center response, Ron came with representatives to the lab to handle and record where it was found to experience. their agencies’ special requests. determine the cause of the accident. These tasks were compounded by the He called an interagency meeting The GIS lab served multiple purpos- immensity of the debris field—more among FEMA, the National Aero- es. A phone bank took public calls than a million acres (400,000 ha) nautics and Space Administration when people found possible shuttle (NASA), and the TFS. “Interagency material, and TFS personnel entered From the beginning, it was apparent coordination was the key. It could the information into a database and that technology would play a vital have been easy for several entities to sent it to the GIS lab. An EPA scout role in tracking the shuttle material be operating in their own bubble.” team then arrived at the site. Using and mapping the searches. Images a personal digital assistant device, from weather radars and satellites Gathering the the team recorded the latitude and on the morning of the accident, Information longitude of the material. Then the February 1, 2003, helped determine Because GIS played such a piv- material was labeled and collected. the debris field. To map and coor- otal role, it became an entity unto Every evening, the team downloaded dinate the information, searchers itself, although it still technically the information into a database. quickly turned to geographic infor- stayed under the planning section. When field searchers found shuttle mation systems (GISs). “Because of the scope of operation material, they flagged it and an EPA and the integration of GIS into the recovery team followed the same On February 2, Greg Atwood, district management of the recovery effort,” procedure in retrieving the mate- forester for the Texas Forest Service said TFS Fire Risk Assessment rial. Maps were then made with the (TFS), was one of the first to respond Coordinator Tom Spencer, “GIS was collected information to delineate to the disaster field office (DFO) in much more than a technical spe- where the shuttle material was Lufkin, TX. He was armed with a cialty on this incident.” found. general knowledge of GISs and a desire to help.

Traci Weaver is a fire prevention specialist for the Texas Forest Service, Granbury, TX.

Volume 65 • No. 2 • Spring 2005 35 Devising a Standard “The maps were considered a matter were still recorded on the official of national security,” said Incident map at the DFO for the archive. Formatting the maps was controver- Commander Mark Stanford. “When sial throughout the incident. “The there are sensitive issues involved— information coming from others By the end of the response, 115 as there always are with shuttle in the field was being delivered in people from across the country had launches—you have to work under a multitude of formats,” Langhelm worked in the GIS lab, an average of somewhat rigid requirements.” said. “A standard for data collection 35 people a month. The lab produced was necessary in our evolving data 30,000 maps, sometimes as many as Some maps were archived for the collection plan.” 1,000 in a day, making it the largest historical record and the rest shred- GIS effort to date. ded. Initially, maps were banned “The first ground search crews from public viewing. They were off requested that the maps be format- “There were problems, but we limits for use in media briefings ted in decimal-degrees to match the worked pretty aggressively to resolve and were not allowed for posting in settings of their [global positioning any issues that would arise day-to- public areas at the base camps until system] units, which would make day, like connectivity or firewall cleared by NASA. it easier for them to place items on problems,” Langhelm said. “But the the map,” said Spencer, who served success part of this incident and GIS “We learned that it is critical to be as the link between the field and the operation was incredible. To bring very clear upfront about any restric- DFO for GIS products. “We agreed. 115 people together, most of whom tions regarding GIS products, as well As other crews and teams arrived had never met, and get everyone as the importance of proper handling to the incident, some requested in there working together to meet of sensitive items,” Spencer said. that the maps be changed back to the objective is pretty amazing. In “Maintaining the integrity and prop- degrees, minutes, [and] seconds. a peacetime environment, it’s easy er use of the lab products had to be Since we had already made the deci- for people to get frustrated and walk a total team effort between the field sion to go with decimal-degrees, we away. We don’t do that in a disaster and the lab. There were some con- decided to keep this format through- mode.” cerns along the way, but the teams out the incident.” did an outstanding job of adapting to Langhelm said that every disaster the demands of this response.” “It was an important lesson learned,” is different, requiring responders to he added. “Make sure you include as adapt to an ever-changing environ- many key people as possible when ment. “The shuttle was a completely you make the initial decision, then “Interagency coordination different scenario than 9/11,” he clearly communicate to everyone was the key. It could said. “This was such a huge area, but why those choices were made.” have been easy for the core management of the incident was quite small. In New York, it was several entities to be Grid Maps the opposite.” operating in their own Providing grid maps for such a large “Interagency coordination is key, geographical area was challenging. bubble.” beyond that it’s very dynamic,” For phase two of the incident, four –Ron Langhelm, FEMA Langhelm said. “For the most part, incident management teams were GIS people supporting wildfire have brought in and set up in Corsicana, that market cornered. It’s when a Hemphill, Nacogdoches, and By mid-March, NASA had relaxed new event crops up that it’s chal- Palestine. Even though the camps its restrictions on the incident lenging. Having the ability to utilize were spread across 190 miles (306 maps. While some of the teams were your knowledge base and technology km) of East Texas, search managers cleared to make their own maps for to adapt to a new environment is made the decision to keep all the tactical purposes, the search data critical.” ■ GIS mapping at the DFO.

Fire Management Today 36 USING SATELLITE IMAGERY FOR BURNED AREA EMERGENCY RESPONSE

Andrew Orlemann, Jess Clark, Annette Parsons, and Keith Lannom

urned area emergency response (BAER) teams are BAER team members need maps of burn severity Bamong the first onsite spe- within a few days after a fire so they can quickly cialists to visit a postfire area. BAER focus on mapping ecological and safety concerns. team members need burn severity maps within a few days of any fire so they can quickly focus their field In 1996, the USDA Forest Service’s RSAC Products and time on assessing and mapping eco- Remote Sensing Applications Center Services logical and safety concerns. (RSAC), in conjunction with several The basic product RSAC provides to Forest Service hydrologists and soil field-level BAER team members is a Traditionally, burn severity maps scientists, tested an airborne color satellite image (fig. 2). Whether in were “quick-and-dirty” jobs made infrared digital camera for its ability digital or paper form, the image pro- from helicopters hovering over a to quickly record postfire conditions. vides a synoptic view of the burned burn. The mapper hand-drew large, The resulting image mosaic proved area and allows for consistent map- imprecise polygons on a topographic useful and led to further technique ping decisions. In addition, the base map (fig. 1). Aerial observations refinement and eventually com- image is often used to provide infor- were then supplemented with a cur- mercialization. Today, BAER team mation to the public. sory drive-through. leaders hire contractors to fly over and build a digital image of a burned But times have changed. Through area. remote sensing, BAER teams now have technologically advanced maps, Although an excellent source of enabling them to quickly access the information, the process is often information they need. technologically cumbersome. The imagery must be a mosaic that is BAER Teams Go Digital reprocessed before use. Availability For more than 30 years, BAER teams is also an issue, especially in areas have conducted rapid assessments without contractors or aircraft to and identified potential flood and gather the data. As a result, RSAC sediment source areas that could began looking for other options. potentially pose threats to down- stream life, property, and resources. In 2001, RSAC conducted a pilot When threats are identified, the project to provide BAER teams teams pinpoint opportunities for with moderate-resolution satel- remediation. lite imagery and derived products. The pilot project was so successful Andrew Orlemann, formerly lead analyst for that RSAC acquired funding from the BAER mapping program at the USDA the National Fire Plan through the Forest Service’s Remote Sensing Applications Forest Service’s Watershed and Air Center (RSAC), is the environmental coor- Management Staff in Washington, Figure 1—Example of an early BAER burn dinator for the Escalante Ranger District, severity map. These maps were often created Dixie National Forest, Cedar City, UT; Jess DC, to continue the program into by drafting rough polygons on forest visitor Clark and Annette Parsons are remote 2002. or topographic maps. sensing/GIS analysts for RSAC; and Keith Lannom, the former operations program leader for RSAC, is now the district ranger on the Tellico Ranger District, Cherokee National Forest, Telico Plains, TN.

Volume 65 • No. 2 • Spring 2005 37 Examples From 2002 Hayman Fire. The Hayman Fire burned from June 8 to July 2, 2002. At 138,000 acres (56,000 ha) it was the largest fire in Colorado history, and it affected critical water supplies for Colorado’s booming Front Range communities. The BAER team was under pressure to quickly create a high-quality burn severity map so treatments could begin.

On June 16, 2002, RSAC acquired a Landsat image that was only mar- ginally useful, because it was very cloudy and the fire was still growing. Finally, on June 23, RSAC obtained Figure 2—Landsat 7 ETM+ image of the Cannon Fire, June 22, 2002. Burned areas appear red, healthy vegetation appears green, and bare ground or light grasses appear pink. virtually smoke-free imagery from both SPOT and Landsat. BARC map- If the image is cloud and smoke free, • Create three-dimensional models ping efforts began, and the BAER RSAC will provide a burned area of the burned area, and burn severity map was ready days reflectance classification (BARC) • Prepare graphics to publicize later. map. The map is typically created areas with increased erosion using change detection on band potential. Biscuit Fire. Biscuit was huge, ratios, such as the Normalized Burn burning from mid-July 2002 into Ratio or the Normalized Difference the first week of September and cov- Vegetation Index. The BARC output ering nearly 500,000 acres (200,000 Feedback was positive: ha) on the Siskiyou National Forest is delivered to the field in two forms. Like Biscuit, Rodeo– One has four classes—high, moder- in southwestern Oregon. As early as ate, low, and unburned. The other Chediski was too big to August 15, RSAC obtained a Landsat has 255 discrete values that allow effectively map in any image and produced a BARC map for field users to make their own clas- other way. the initial BAER team, but the image sification decisions. was partly smoke obscured. On August 30, RSAC obtained a nearly smoke-free Landsat image (fig. 3) Digital elevation models and derived Since its inception, the RSAC BAER and issued a new BARC map on products, such as slope, elevation, support team has provided emergen- August 31. After several helicopter and aspect, are available for each cy satellite imagery acquisition and reconnaissance flights and a few days BAER team. In addition, if a BAER burn severity mapping services to of editing, the BAER burn severity team requests, RSAC uses the digital more than 160 BAER teams. RSAC map was finalized. elevation models to create and deliv- has purchased more than 260 satel- er three-dimensional visualization lite images and delivered imagery The feedback from the Biscuit BAER tools, such as fly-throughs. and derived products to BAER teams team was positive. For a fire of this within 12 to 24 hours after satellite size, the traditional methods of map- BAER teams use RSAC products to: acquisition. ping via helicopter or road overlook were impractical. Other options, • Focus the team on areas of RSAC also provides technical sup- including high-resolution fixed wing greatest concern, port. Specialists work with BAER mosaics or IKONOS images, were • Develop a BAER burn severity team members to help them under- costly and time consuming. A 90- map, stand and use the furnished prod- foot (30-m) Landsat-derived BARC • Help predict runoff response in ucts. They provide both onsite and hydrologic models, online training and are prepared to join BAER teams when requested.

Fire Management Today 38 map proved its worth, and the BAER 27—1 day before containment—was end, RSAC, along with its partners team was able to complete its work a smoke-free image of the entire in Forest Service Research and the in record time. burned area. U.S. Department of the Interior U.S. Geological Survey, is building satel- Rodeo Fire. Rodeo–Chediski was The fire threatened a number lite-based tools and products to help nearly as big as Biscuit, covering of small towns, along with rare rehabilitation program managers about 460,000 acres (184,000 ha) sequoia groves on the Giant Sequoia with effectiveness monitoring. For and threatening numerous small National Monument. RSAC special- example, RSAC worked in 2004 with towns around the Apache–Sitgreaves ists worked with BAER team mem- rehabilitation staff on the Hayman National Forest in Arizona. Soon bers to refine BARC map production (2002) and Cerro Grande (2000) after it started in mid-June, a BAER for that particular ecosystem. In the Fires to provide satellite-based veg- team began making field observa- end, Sequoia National Forest person- etation recovery estimates. tions, but afternoon clouds building nel called RSAC’s satellite support a along the Mogollon Rim effectively major factor in the McNally BAER For more information on RSAC or prevented the acquisition of satellite team’s success. the BAER support program, visit our imagery. Website at . ■ Finally, on July 7, 2002, the day the The Remote Sensing fire was contained, RSAC obtained Application Center has mostly cloud-free coverage from both Landsat and SPOT. Feedback provided emergency was again positive: Like Biscuit, satellite imagery Rodeo–Chediski was too big to acquisition and burn effectively map in any other way. severity mapping Satellite-based postfire condition maps saved the BAER team days, services to more than perhaps weeks, of helicopter sketch- 160 BAER teams. mapping. Other agencies acquired high-resolution images of the Rodeo Fire from a commercial vendor, but The Future large file size dramatically slowed Past success has shown that satellite processing. data can facilitate the quick evalua- tion of postfire effects. Remote sens- McNally Fire. The McNally Fire ing technologies also promise to play started on July 21 near Kernville, a growing role in long-term BAER CA. By August 28, it had burned treatment monitoring. Recent stud- more than 150,000 acres (60,000 ies by the General Accounting Office, ha) on the Sequoia National Forest. among others, have pointed to the Figure 3—Landsat 7 ETM+ image of the The BAER team began assembling Biscuit Fire, August 30, 2002. The Biscuit lack of monitoring as a problem that Fire burned about 500,000 acres (200,000 early, and RSAC collected four satel- forest managers must begin address- ha) on the Siskiyou National Forest in lite images of the burned area, but ing in a systematic manner. To that southern Oregon and northern California. only the fourth, acquired on August This image is displayed using band combina- tion RGB = 7,4,3.

Volume 65 • No. 2 • Spring 2005 39 SATELLITE MAPPING OF WILDLAND FIRE ACTIVITY

Keith Lannom, Brad Quayle, and Mark Finco

he USDA Forest Service’s Remote Sensing Applications The MODIS sensor is well suited to the broad-scale TCenter (RSAC) uses satellite mapping and monitoring of wildland fires. imagery to detect active fire loca- tions and produce national- and regional-scale maps of wildland fire sensor onboard. With the addition time using image data collected by activity (fig. 1). Maps and other geo- of Aqua, the number of images RSAC’s receiving station. UMD and spatial data on the Website provide a acquired each day by a MODIS GSFC concurrently developed a sys- spatial component for the National instrument doubled. tem for detecting fires worldwide Interagency Fire Center (NIFC) using MODIS imagery from the Incident Management Situation (Sit) Each MODIS sensor contains several National Oceanic and Atmospheric Report. The active fire maps provide thermal channels for identifying Administration’s Near-Real-Time information on a fire’s geographic fire activity, a footprint 1,448 miles Processing System. RSAC uses data extent and recent activity, whereas (2,330 km) wide that provides a from this system to obtain infor- the Sit Report lists large fires, synoptic view of large areas, and an mation about fires in the Eastern resources allocated, estimated con- orbit configuration that allows each United States and Alaska, which are tainment date, and costs to date. sensor to collect thermal image data locations currently beyond the range over middle to high latitudes at least of RSAC’s receiving station. Satellite Fire Detection twice daily. In December 1999, the National Aeronautics and Space Adminis- Terra and Aqua satellites broadcast tration (NASA) launched Terra, imagery data collected by each the first of a new generation of MODIS sensor in real time. Direct Earth Observation System satel- broadcast stations receive the data in lites. Onboard the Terra satellite was real time, which then permits near- the Moderate Resolution Imaging real-time imagery analysis. Spectroradiometer (MODIS) sensor, designed for global-scale monitor- MODIS Rapid Response ing of the land, oceans, and lower In the spring of 2001, RSAC, NASA’s atmosphere. MODIS is well suited to Goddard Space Flight Center broad-scale mapping and monitoring (GSFC), and the University of of wildland fires because of its ther- Maryland (UMD) established MODIS mal imaging capabilities, wide field Rapid Response—a program that of view, and multiple daily exploits the capabilities of MODIS overpasses. imagery to provide spatial informa- tion about wildland fires. UMD and In May 2002, NASA launched a sec- GSFC developed algorithms that ond satellite, Aqua, with a MODIS analyze MODIS imagery to detect active fire locations. RSAC operates a Figure 1—Example of an active-fire map Keith Lannom, the former operations pro- MODIS receiving station in Salt Lake from October 27, 2003. Several wildfires gram leader for the USDA Forest Service's occurred concurrently in wildland/urban Remote Sensing Applications Center (RSAC), City, UT (fig. 2). interface areas of southern California in late Salt Lake City, UT, is the district ranger October 2003. The red areas represent loca- on the Telico Ranger District, Cherokee MODIS Rapid Response makes tions with fire activity within the previous 24 National Forest, Telico Plains, TN; Brad hours, and the yellow areas show locations Quayle is the remote sensing/GIS specialist it possible to detect fires in the with fire activity before that. and is the senior scientist, Red Castle Western United States in near-real Resources.

Fire Management Today 40 The Aqua and Terra MODIS fire detection data from both facilities are combined and mapped on the basis of topography, State bound- aries, urban locations, and major road systems to create active fire maps. Active fire maps, produced since July 4, 2001, are developed for several geographic areas that cover the entire United States and are provided in Adobe Acrobat PDF and JPG formats. The maps are updated several times daily at strategically designated times to coincide with Aqua and Terra MODIS passes. They are posted on the MODIS Active Fire Figure 2—The Remote Sensing Applications Map Website at . from the Terra and Aqua satellites six to eight times daily. Figure 3—MODIS imagery of northern Utah acquired on August 12, 2001. The Great Salt The MODIS sensor Lake is visible on the left. Active Fire Maps collects imagery at any The MODIS sensor collects imagery location in the middle Fire researchers and GIS special- at any location in the middle to high to high latitudes of the ists also benefit from the MODIS latitudes of the Earth at least twice Earth at least twice daily. Rapid Response when they integrate daily. Some of the extreme northern fire point geospatial data into their latitudes are captured four to six research or mapping activities. times a day. For example, Terra col- The active fire detection GIS geo- lects imagery over the contiguous 48 Applications of Active spatial data files are available from States at approximately midday and Fire Maps the Forest Service Geospatial Data midnight. Aqua captures the same The purpose of this project was to Clearinghouse at . The 2004 file is updated perspective on the current wildland hourly, the 2002 and 2003 files are When either the Terra or Aqua fire situation for wildland fire man- for each year, and the 2001 file cov- satellites pass over the Western agers. Several active fire map users ers July 4, 2001, to December 31, United States, the RSAC receiv- have developed some creative map- 2001. ing station tracks the satellite and ping applications. A county manage- acquires MODIS data as they are ment agency in Oregon downloaded Thanks to the collaborative efforts broadcast from the satellite (fig. the map for their county each day of RSAC, GSFC, and the UMD, the 3). The acquired data are compiled and posted it on the county Website MODIS active fire maps provide a and further processed to identify to keep the residents informed about rapid, consistent, and easily acces- active fire locations for the Western local fire activity. Incident informa- sible source of fire mapping for the United States, Canada, and Mexico. tion staff have selected the area United States. Active fire maps are GSFC performs similar operations to around their fire and posted it on used with the NIFC Sit Report as acquire imagery and detect fires on the incident’s Website. Using active well as for strategic planning by the the East Coast and in Alaska. fire maps, RSAC fulfilled an e-mail interagency fire community and for request from a tourist who was plan- general public information. ■ ning a vacation and wanted to avoid active fire areas.

Volume 65 • No. 2 • Spring 2005 41 GUIDELINES FOR CONTRIBUTORS Editorial Policy title(s), affiliation(s), and address(es) of the etc.). At the end of the manuscript, include Fire Management Today (FMT) is an inter- author(s), as well as telephone and fax num- clear, thorough figure and photo captions national quarterly magazine for the wildland bers and e-mail information. If the same labeled in the same way as the correspond- fire community. FMT welcomes unsolicited or a similar manuscript is being submitted ing material (figure 1, 2, 3; photograph A, B, manuscripts from readers on any subject elsewhere, include that information also. C; etc.). Captions should make photos and related to fire management. Because space is Authors who are affiliated should submit a illustrations understandable without reading a consideration, long manuscripts might be camera-ready logo for their agency, institu- the text. For photos, indicate the name and abridged by the editor, subject to approval by tion, or organization. affiliation of the photographer and the year the author; FMT does print short pieces of the photo was taken. interest to readers. Style. Authors are responsible for using wildland fire terminology that conforms Electronic Files. See special mail- Submission Guidelines to the latest standards set by the National ing instructions above. Please label all Submit manuscripts to the general Wildfire Coordinating Group under the disks carefully with name(s) of file(s) and manager at: National Interagency Incident Management system(s) used. If the manuscript is word- System. FMT uses the spelling, capitaliza- processed, please submit a 3-1/2 inch, IBM- USDA Forest Service tion, hyphenation, and other styles recom- compatible disk together with the paper Attn: Melissa Frey, F&AM Staff mended in the United States Government copy (see above) as an electronic file in one Mail Stop 1107 Printing Office Style Manual, as required by of these formats: WordPerfect 5.1 for DOS; 1400 Independence Avenue, SW the U.S. Department of Agriculture. Authors WordPerfect 7.0 or earlier for Windows 95; Washington, DC 20250-1107 should use the U.S. system of weight and Microsoft Word 6.0 or earlier for Windows tel. 202-205-0955, fax 202-205-1401 measure, with equivalent values in the 95; Rich Text format; or ASCII. Digital pho- e-mail: [email protected] metric system. Try to keep titles concise tos may be submitted but must be at least and descriptive; subheadings and bulleted 300 dpi and accompanied by a high-resolu- Mailing Disks. Do not mail disks with material are useful and help readability. As a tion (preferably laser) printout for edito- electronic files to the above address, because general rule of clear writing, use the active rial review and quality control during the mail will be irradiated and the disks could be voice (e.g., write, “Fire managers know…” printing process. Do not embed illustrations rendered inoperable. Send electronic files by and not, “It is known…”). Provide spellouts (such as maps, charts, and graphs) in the e-mail or by courier service to the managing for all abbreviations. Consult recent issues electronic file for the manuscript. Instead, editor at: (on the World Wide Web at

Fire Management Today 42 PHOTO CONTEST ANNOUNCEMENT

Fire Management Today (FMT) invites Rules • A panel of judges with photography you to submit your best fire-related • The contest is open to everyone. You and publishing experience determines images to be judged in our annual com- may submit an unlimited number of the winners. Its decision is final. petition. Judging begins after the first entries taken at any time. No photos • We will eliminate photos from com- Friday in March of each year. judged in previous FMT contests may petition if they are obtained by illegal be entered. or unauthorized access to restricted AWARDS • You must have the right to grant the areas; lack detailed captions; have date All contestants will receive a CD with Forest Service unlimited use of the stamps; show unsafe firefighting prac- the images and captions (as submitted) image, and you must agree that the tices (unless that is their express pur- remaining after technical and safety image will go into the public domain. pose); or are of low technical quality reviews. Winning images will appear in Moreover, the image must not have (for example, have soft focus or show a future issue of FMT and will be pub- been previously published. camera movement). licly displayed at the Forest Service’s • We prefer original slides or negatives; • You must complete and sign a release national office in Washington, DC. however, we will accept duplicate statement granting the USDA Forest slides or high-quality prints (for exam- Service rights to use your image(s). Winners in each category will receive: ple, those with good focus, contrast Mail your completed release with your level, and depth of field). Note: We entry or fax it to 970-295-5815 at the • 1st place—Camera equipment worth will not return your slides, negatives, same time you e-mail your digital- $300 and a 20- by 24-inch framed or prints. image files. copy of your image. • We will also accept digital images if • 2nd place—A 16- by 20-inch framed the image was shot at the highest Mail entries to: copy of your image. resolution using a camera with at USDA Forest Service • 3rd place—An 11- by 14-inch framed least 2.5 megapixels or if the image Fire Management Today Photo Contest copy of your image. was scanned at 300 lines per inch or Madelyn Dillon • Honorable Mention—An 8- by 10-inch equivalent with a minimum output 2150 Centre Avenue framed copy of your image. size of 5 × 7. Digital-image files should Building A, Suite 361 be TIFFs or highest quality JPGs. Fort Collins, CO 80526 Categories • You must indicate only one or • Wildland fire competition category per image. To e-mail images and captions to: • Prescribed fire ensure fair evaluation, we reserve [email protected] and • Wildland/urban interface fire the right to change the competition fax signed release form to • Aerial resources category for your image. 970-295-5815 (attn: Madelyn Dillon) • Ground resources • You must provide a detailed caption • Miscellaneous (fire effects; fire weath- for each image. For example: Postmark Deadline er; fire-dependent communities or A Sikorsky S–64 Skycrane delivers First Friday in March species; etc.) retardant on the 1996 Clark Peak Fire, Coronado National Forest, AZ. Photo: name, professional affiliation, town, state, year image captured.

SAMPLE PHOTO RELEASE STATEMENT

Enclosed is/are ______(number) image(s) for publication by the USDA Forest Service. For each image submitted, the contest category is indicated and a detailed caption is enclosed. I have the authority to give permission to the Forest Service to publish the enclosed image(s) and am aware that, if used, it/they will be in the public domain and appear on the World Wide Web.

Contact information:

Name ______

Institutional affiliation, if any ______

Home or business address ______

______

Telephone number ______E-mail address ______

Volume 65 • No. 2 • Spring 2005 43 Fire Management Today 44