Historic, Archive Document Do not assume content reflects current scientific knowledge, policies, or practices. I'- S9 6^ i( (Ji^-^ • A Mann United States ^ Gulch Department QxUjif Fire: ofAgriculture r^*-'^^ ForestService Race That Couldn't Intermountain Research Station Won Be General Technical Report INT-299 May 1993 v3 Richard C. Rothermel u4s( THE AUTHOR work unit from 1966 until 1979 and was project leader of the Fire Behavior research work unit until 1992. RICHARD C. ROTHERMEL is a research physical sci- entist stationed at the Intermountain Fire Sciences Labo- RESEARCH SUMMARY ratory in Missoula, MT. Rothermel received his B.S. degree in aeronautical engineering at the University of The Mann Gulch fire, which overran 16 firefighters Washington in 1953. He served in the U.S. Air Force in 1949, is analyzed to show its probable movement as a special weapons aircraft development officer from with respect to the crew. The firefighters were smoke- 1953 to 1955. Upon his discharge he was employed at jumpers who had parachuted nearthe fire on August 5, Douglas Aircraft as a designer and troubleshooter in 1949. While they were moving to a safer location, the the armament group. From 1957 to 1961 Rothermel fire blocked their route. Three survived, the foreman was employed by the General Electric Co. in the aircraft who ignited an escape fire into which he tried to move nuclear propulsion department at the National Reactor hiscrew, and two firefighterswho found a route to safety. Testing Station in Idaho. In 1961 Rothermel joined the Considerable controversy has centered around the Intermountain Fire Sciences Laboratory (formerly the probable behaviorof the fire and the actions ofthe crew Northern Forest Fire Laboratory), where he has been members and theirforeman. Modern safety techniques engaged in research on the behavior of fire. He re- used by 73 firefighters who escaped injury after being ceived his master's degree in mechanical engineering trapped on the Butte Fire in 1985 are described for at the Colorado State University, Fort Collins, in 1971. comparison. Hewas project leaderofthe Fire Fundamentals research Intermountain Research Station 324 25th Street Ogden, UT84401 — Mann Gulch A Race That Fire: Won Couldn't Be Richard C. Rothermel INTRODUCTION This paperisnotintendedtobe acomplete account ofthe events surroundingthe Mann Gulch fire OnAugust 5, 1949, a wildfire overran 16 firefight- Macleanprovides such anaccountin hisbook, "Young ers in Mann Gulch on the Hel—ena National Forest in MenandFire." Rather, this paperexaminesthe prob- Montana. Onlythree survived the foreman andtwo ablebehaviorofthe fire andthe movements ofthecrew members ofan 18-man smokejumper crewthat had duringthelast20 minofthetragedy. The analysisis parachuted into a small valley or gulch nearthe fire. a reconstruction ofwhat probably happened. Even These deaths were a shockingloss to the firefighters' though events late inthe story appear to be worked families andfriends. The tragedy was also a severe out precisely, they cannot be verified and must be — blow to the Forest Service, which hadnot experienced taken for whatthey are reconstructed estimates. a fatality during a decade ofsmokejumping andwas TheMannGvilchfire had serious consequences for extremelyproud ofits ehte firefighters. Repercussions the Forest Service anditsresearchbranch. While in from this incident were severe and long lasting. Mann Gulch to investigate the fire duringthe faU of In 1979, some 38 years after the Mann Gulch fire, 1949, Harry Gisborne, the pioneerfire scientistinthe author Norman Maclean contacted the Intermoun- NorthernRockies, suffered a heart attack and died. tain Fire Sciences Laboratory (IFSL) to help clarify Jack Barrows, who succeededGisborne, was directed the fire's behavior for abook he was writing about to expedite research on fire behavior. He championed thetragedy. Maclean, who diedin 1990, iswell known modem scientific research, estabhshingthe Northern as author of"A River Runs Through It." His second ForestFire Laboratory (now the Intermountain Fire andfinalbook, "YoimgMen andFire,"was published Sciences Laboratory) at Missoulain 1960. in 1992. Macleantaught attheUniversity ofChicago, but spent his summers at Seeley Lake, MT, near his ANALYZING FIRE BEHAVIOR boyhood home in Missoula. His researchuncovered conflictingstories ofhowthefirehad overrunthefire- Analysis offirebehaviorrequires dataonfuels, fighters. When Maclean learnedthat FrankAlbini weather, topography, andthefire situation. Fortunate- and I were developingmethods to predict fire behav- ly, Laird Robinson, a former smokejumper, teamed ior, he asked ifwe could use the same methods to re- up withMacleanto learn as much as possible about construct the behavior ofthe fire that caughtthe fire- the Mann Gulch fire. They persuaded the two living fighters at Mann Gulch. crew members, Robert Sallee and Walter Rumsey, to Initially, I was reluctant to complywith his request jointhem atMann Gulchtoreconstructthe final min- because ofthe controversy surroundingthe fire and utes ofthe tragedy. This information, coupled with my desire notto reopen emotional wounds. Butthe theirsearchofthe archives andRobinson's knowledge Forest Service, U.S. Department ofAgriculture, Fire ofthe site, provided the datausedto reconstruct the Laboratory atMissoiolawas conceivedinthe aftermath fire behavior andthe movements ofthe crew. ofthe Mann Gulch fire. It seemed only right that we Forthose readers who wonder how itis evenre- use the results ofovirresearchtohelp explainthe com- motelypossibletoreconstructthese events, fivetypes plicatedinteractions offuels, weather, andtopography ofinformation make it possible: that produced such a tragedy. Several questions about the fire had neverbeen 1. Thelocationofthe crew'smovements and actions were recorded; the distances between significant ac- resolved: tions were checked and measuredby Maclean and • How did the fire, whichwas burning on a ridge Robinson. when the crewlanded, get belowthem atthe mouth 2. The crew's foreman survived; his testimony dur- ofthe gulch? ingthe initial investigation provided estimates ofthe • What was their position relative to the fire as it fire's positionwithrespect to the crew at significant progressed? moments. Why • couldn't they escape? 3. Archivedweather data and site maps were • Did the escape fire ignitedbythe crew foreman available. overtake his own crew? 1 4. Survivors Sallee and Rumseyrecalledthetypes ofthe fire. Thejump, completed between 3:50 and ofvegetation cover they were moving through. 4:10 p.m., was considered routine by the foreman, 5. Calculations ofthe fire's approximate rate of R. Wagner Dodge, andthejvmipmaster. Earl Cooley. spread along sections ofthe route were integrated The fire's size and location and the area where the withknovm distances andtimes to estimate the rate jiunpers landed are shown infigure 1. The cargo drop at which the crew traveled. did not go smoothly; the plane, a twin engine DC-3, The fire's rate ofspread, its intensity, andits flame encovmteredheavyturbulence at normal drop altitude length were calculatedby usingmathematicalmodels andwas forcedto climb before droppingthe remaining developed from a combination oflaboratory fire ex- cargo. Fire-fightinggearwas scattered andthe crew's periments and field data that were programmed for radio was broken. By the time thejumpers gathered use by fire analysts in the field (Rothermel 1983). their gear, it was nearly 5 p.m. They did notfeel the The initial calculations, made with the fire behavior firethreatenedthemthen. Survivingcrewmember chip developed for the TI-59 calculator(Burgan 1979), Salleestated intheofiBcialreportoftheincident: were verifiedwiththe BEHAVE computer program I tookalookatthe fire and decideditwasn'tbad. (Andrews 1986). Itwasburningon top ofthe ridge and I thoughtit would continue on up the ridge. I thoughtitprob- THE SCENARIO ablywouldn'tbummuchmore that nightbecause itwas the endoftheburningperiod (forthat day) The Mann Gulch fire was spotted at 12:25 p.m. on anditlookedlikeitwouldhavetoburn downacross August 5, 1949, a very hot and windy day. The fire alittle saddle beforeitwentuphill anymore. was in the Gates ofthe Mountains WildArea (now While getting the crew and equipment organized, the Gates ofthe Mountains Wilderness)just east of foreman Dodge heard someone shoutingnearthe fire. the Missouri River, 20 mi north ofHelena, MT. Tem- He instructedsquadleaderWilliam Hellmanto equip peratures thatdayreached 97 °F in Helena. The fire the crew andleadthem down the north side ofMann started near the top ofa ridge between Mann Gulch Gulch toward the Missouri River while Dodge went and Meriwether Canyon. Mann Gulch is a minor on ahead to see who was shouting. Near the top of drainage, leading into the Missouri River from the the ridge closetothe headofthe fire. Dodge metJames east. Itis fiinnel shaped, narrowingto awidth ofone- Harrison, arecreation and fire preventionguard from fourth mi at the river. The highest flanking ridge, the HelenaNational Forest. Harrison, based atnearby wherethe fire started, is onthe southside ofthe drain- Meriwether Campground, had been the first to spot age between Mann Gulch and Meriwether Canyon. the fire and had been tryingto keep it from crossing The ridge on the north side ofthe drainage, where into Meriwether Canyon while he awaited help. The the fire overranthe crew, is not as high astheridge fire was still movingnortheast alongthe ridge between to the south. Mann Gulch and Meriwether Canyon. Dodge decided Vegetation on the north side ofMann Gulch was the ridge was not the safest place to attack the fire, mature 60- to more than 100-yr-old ponderosa pine. so he and Harrison did not stay there. Afi;er a quick The south side was covered with 15- to 50-yr-old lunchtheycaughtup withHellman andthe crewwho Douglas-fir, mixed with mature ponderosa pine and were traversingthe slope, headingdovmthe gulchto- some maturejuniper. Frontingthe riverwas a stand wardtheMissouri River. The smokejumpers behoved of60- to more than80-yr-old Douglas-fir. Mixed pine they were goingto attack the fire in another location, and fir grew in the bottom ofthe gulch. certainlya saferlocationthatwould be on the upwind A distinctmoisture gradientis evidentinthe gulch; side ofthe fire, near the river. the lowerslopes receive more moisture thanthe upper Dodge had aclearviewofthe fire and could see it slopes. This influences the vegetation found in the was burning more rapidly than before. In Helena, understory beneath the forest canopy. At the time the vnnd had been blowing from the north and east ofthe fire, lowerelevations had heavierundergrowth, at 6 to 8 mi/h that afternoon. At 3:30 p.m. the wind which gave way to scattered timber and grass in the switched tothe south, increasedto 24 mi/h, and contin- drier areas farther up the gulch. uedto blow strongly from the south at 14 to 22 mi/h. Access to this roadless area is difficult. Therefore, Because ofthe orientation ofthe canyons and ridges, smokejumpers were called when the fire was discov- a strong southerly wind would create extreme turbu- ered. One ofthebasictenets offire fightingis toreach lence at the mouth ofMann Gulch, producing strong afire quickly. Then it can be attackedwhile itis still wands thatwould blow upthegulchtoward the crew. small. Smokejumpers are very effective at reaching At about 5 p.m. Canyon Ferry District Ranger Robert afire quicklybecause theytravel by airplane anduse Jansson had reached the mouth ofMann Gulch by boat parachutes to land near the fire. and was attemptingto walk up the gulch to reach the In this instance, the smokejumper crew was dis- smokejumpers. He estimatedthe wdnd atMann Gulch patched fi-om Missoula, MT, a little over 100 mi west to be between 20 and 30 mL^ with gusts to 40 mi/h, 2 Jumperslanded here4:00^ 4horsemen X — Figure 1 Map of Mann Gulch illustrating the movementofthe crew and the position ofthefire as itapproached the crew at points (pt.) 1, 2, and 3. much strongerthan recorded atHelena. Consequently, blew the fire offthe ridge into the mouth ofthe can- as the crew proceeded down the gulch, they were yon. This idea is supported by motion picture films walking into a strong headwind. (now lost) taken by a Forest Service photographer ForemanDodge andrecreationguardHarrisonover- fi:"om the same aircraftthatdropped Dodge's crew. tookthe crew atabout 5:40 p.m. Survivors Bailee and Accordingto Jack Barrows (1980), who viewed these Rumsey both said the crew was not worried about films manytimes, "Theyclearly showedratherintense safety, but Sallee thought he heard Dodge say some- downdrafts." Barrows, who said he visited the site thingaboutthe thicketofponderosa pine and Douglas- with survivors Dodge and Sallee, concluded that the firthey were in being a death trap. thunder cell downdraft was an important factor in They continued down the canyon for another 5 min, throwing firebrands to the mouth ofMann Gulch. covering about 400 yd, when they saw fire blocking Other fire experts suspected whirlwinds may have their route to the river. Firebrands from the main spreadthe fire. Fire whirls anddowndrafts from thim- fire had started spotfiresinthe timberaheadofthem. derstorms orthe fire's convection columncanoccurto- These fires were rapidly becoming more intense. gether. Harry Gisborne, then director offire research Duringdiscussions afterthetragedythatfollowed, at the Northern Rocky Mountain Research Station, fire experts were particularly concernedwithunder- notes inthe margin ofan official reportthat fire whirls standinghowthefire got fi-om theridge high onthe formed at the mouth ofMann Gulch. This very likely south side ofMannGulchtothemouthofthe gulchand could have been the case because the ridge between latertothe north side. Tothis day, twoopinionspersist. Meriwether and Mann Gulch would cause a southerly One is thatdowndrafts ofa smalllocalthunderstorm windto form avortexonthelee side inperfectposition 3 to loft numerous firebrands and carry them tothe turnaround, point 1. Fire behavioris difficult to north side ofMann Gulch. Gisbome later diednear assess whenfires arejustbeginning, butifthe spot the gulchwhile attemptingtoresolve this issue. fires started at about 5:30, the surface fires could The approximate locations ofthe spot fires at 5:30 have taken 10 min or soto spread slowly and develop are shown in figure 1. The crew turned around at into a running crown fire. At 120 ft/min the crown about 5:45 when they found their route to the river fire would spread 400 yd to point 1 in another 10 blocked by the fire below them. On the map, their min, arrivingthere about5:50, or 4 to 6min afterthe turnaroundis designated point 1. In Maclean's book, crewhad turned around and headed back up the the turnaround is designated point 6. gulch. Figure 2 shows the relative position ofthe crew and the fire from the time the spot fires started THE RACE THAT COULDN'T BE WON until the crewwas overrun. The slope ofthe lines on the graphrepresents the rate oftravel. The steeper The spot fires, which had started in heavy surface the line, the faster the rate. fuels, wouldhave become intense, withflames extend- Foreman Dodge saidthe firewas 150to 200yd away inginto the tree crowns and climbingthe tree boles. from point 1 when the crew turned around. The sur- The tree crownswould have caughtfire, andthe strong viving crew members, Rumsey and Sallee, recognized gusty winds would have pushed the fire through the the dangerandquicklymovedup to stay close to Dodge. crowns. The crew could see a convection column of As the crew moved back up the canyon, the timber black smoke from the burningtree crowns between beganto thin. More grass and brush appeared in the them and the Missouri River. The fire at this stage understory. The crewmay nothave recognizedthe con- was burning in a stand ofmuch denser forest than sequences ofthe fuel change. The lighter fuel would they were in farther up the gulch. Fuels in the area have produced a faster spreading fire. Other factors where the spot fires started were estimated to have were also in the fire's favor. The fire was burningup- been timber litter and live understory vegetation in- hillwith afollowingwind;theuphill grade slowedthe terspersed with accumulations ofheavy dead anddown crew, but it caused the fire to accelerate. woody material. For purposes ofcalculatingfire be- Thecrewcontinued hurrying across the slope along havior in the surface fuel, I estimated the fuel to be an 18 percent uphill grade. The survivors reported about equally divided between fuel models 10 and 12 they traveled through tall grass much ofthe time. (Anderson 1982). The moisture content offine dead Grass would have been fully cured (dried out) by fuels diiringthe hottest partofthe daywas calculated August 5 at this low elevation (3,600 ft). Fuel model from thetemperatiire andhumiditytohavebeen about 3, which represents cured tall grass, was usedto cal- 3.5 percent; the heavier fuels would not have been culate fire behaviox' from the point where the crew quite so dry. Although winds were strong (30 mi/h) turned around (point 1) to the point where they were inthe open, the forestcanopy wouldreduce windspeed told to drop their heavy tools (point 2). underthetrees. Usingguides developedforfire behav- As the timberthinned, the fire would have been ior analysts (Albini and Baughman 1979; Rothermel more exposed to the wind. We estimated midflame 1983), the winds were estimated to have been about winds increased from 5 to about 7.5 mi/h in the thin- 5 mi/h at the midflame height ofthe fire in surface ner timber. We assumed the wind, although gusty fuels (from 3!/2to 5^ft). The slope was estimatedtobe and variable, was blowing in the same direction the 44 percent in the area where the spot fires started. crew was hiurying. The fire's rate ofspread through The spread ofthe fire is derivedfirstfrom the behav- the grass would have been about 170 to 280 ft/min, iorofthe surface fire. Underthese conditions spot considerably fasterthanthe 120 ft/minrate ofspread fires in the surface fuels would have produced an in- as the fire approached point 1. In the grass, flame tense butrelatively slow-movingfire, with an average lengths would have reached 16 to 20 ft. Fireline in- spreadrate ofabout 20 ft/min and flame lengths rang- tensity wouldhave been 2,500to 4,000 Btu/ft s. The ingbetween 7 and 11 ft. The fireline intensity in the fire would still have beenburningthrough the crowns, surface fuels (fuels nearthe ground) wouldhave been but since the trees were more scattered, the surface 400 to 1,000 Btu/fts. Giventhe strongwinds, steep fire was probably moving ahead ofthe crown fire. slope, and unstable atmosphere, the fire would have The distance from point 1 to point 2 is 450 yd, or quicklycrowned as described earlier. Thewind-driven 1,350ft. At aspread rate of170 to 280 ft/min, the fire crown fire burning up the slope would have spread would cover 1,350 ft in 5 to 8 min. Assuming the fire fourto sixtimes fasterthanthe surface fire, or 80to was spotting, the faster rate is probably more appro- 120 ft/min. priate; the fire could have reached point 2 by 5:54. Once the crown fire developed on the steep slope Afterturningaroimd at point 1, the crewwent faster andwasbeing driven bythe turbulentwinds, itwould duringeach successive legofthejourney(fig. 2). But have spread at the faster rate of120 ft/min. Maclean the fire went even faster. estimatesthe spotfires startedabout400ydfrom the 4 4,000 3,500 Leading men (4 horsemen) 3,000 caught 5:57 Ratefrom pt. 2 Slowest men 2,500 caught 5:56 Dodge 2,000 Pt. 3, sets fire 5:55 Jumperforeman (Dodge) and recreation guard 1,500 (Harrison) rejoin the crew 1,000 Fire arrives 500 atpt. 2 5:54 Crewturns 0 around, pt. 1 Spotfires 5:45 -500 start at mouth of Fire arrives t Mann Gulch at pt. 1 5:49 1 5:30 5:35 5:40 5:45 5:50 5:55 6:00 Time — Figure 2 Distance and time graph ofthe estimated positions ofthe crewand the fire. Distances are estimated from the crew's turnaround at point (pt.) 1. The slopes ofthe lines indicate the rate of movement; the steeperthe line, thefasterthe rate. Inhindsight, it's possible to askwhythe crewdidn't Atpoint2, Dodgetoldhis crewtodiscardtheirheavy immediatelybeatthe hastiest possibleretreatas soon tools. Most ofthe crew realized now thatthey were as fire blockedthe route to the river. I suspectthey inrealtrouble. Inhistestimony. Dodge saidthatwhen did notyet realize how serious their situation hadbe- the crew members droppedtheir tools, the fire was come. Whenthey first saw fire blockingtheirroute, only 75 to 100 ydbehindthem. Ifthe fire were trav- the fire was probably in the early stages ofdevelop- eling at 280 ft/min, itwould coverthat distance in mentandhadnotyetbeguntorun. The crewdidnot about 1 min. Therefore, the crewwovddhave reached have atrail to follow on this slope, whichbecame as point 2 about 1 min ahead ofthe fire, at about 5:53. steep as 76 percentleadingto the ridge ontheirleft. The time ittookthem to go fi-om point 1 to point 2 Theywere sidehiUingon aroutewithabout an 18 per- would havebeen about 8 min. centslope. The terrainwas broken andlittered vnth From point 2 to point 3, where Dodge lit the escape loose rock. The river which they had set outto reach fire, conditions were similarto those firom point 1 to was nowbehindthem, andtheremusthave beensome point2, exceptthatthetimberwaseventhinner. This confusion about where they were going. Itwas very allowedthe wind nearthe surface toincrease to an — hot Maclean believes it w—as much hotterthan the estimated9to 13 mi/h. Consequently, the fire's spread 97 °F measured in Helena andinitiallytheywere ratewouldhaveincreasedto somewherebetween 360 still carrying packs andtools. One crew member, and 610 ft/min. Flame lengths in the surface vegeta- Navon, was reported to have photographedthe fire tionwouldhavebeen24to 30 ft, withthe flames fi-om across the canyon, indicatingthey were notyet seri- crowningtreesreachingmuchhigher. Firelineintensi- ouslyconcernedvnthescapingfi-omthefire. Account- ties fi-om the surface fuel alone would have beenbe- ingfor the time ittookthem toturn around and start tween 5,500 and 9,000 Btu/ft-s. upthe gulch, theyaveraged 170ft/minorabout2mi/h The most pernicious effect ofthe crew's retreatup overthis leg oftheirretreat. I suspect they were go- the canyonwas thatthey had moved out ofthe tim- ingfaster as they approached point 2. berinto open areas more exposed to the wind. Carl 5 Wilson (1977) at the Riverside, CA, fire laboratory lowvalue. Probablythe most dangerous influence was foundthatmostfire fatalities have occurred inflashy the windthatmay have beengustingto 40 mi/h in the fuels oron the periphery oflargerfires. Fires spread open. It was probably scouring close to the ground, rapidly in light fuels and can change direction and giving midflame winds of15 to 20 mi/h. With these accelerate quickly as the wind shifts. Ironically, the conditions, the fire would have spread between 600 crew did notreduce theirdangerwhenthey movedinto and 750 ft/min, averaging 660 ft/min, with flames lighter fuels; in fact, the fire would have accelerated ranging from 10 to 40 ft long. The fireline intensity in the lighter, flashier fuels, especially as the surface would have ranged very widely from about 850 to wind became stronger. The acceleration is reflected nearly 16,000 Btu/ft s. The firewould have been char- in the increasingly steep lines offigure 2. acterizedby extremely rapid spread, with pulsing Based on the testimony ofsurvivors Sallee and flames thatwouldrise and fall as the buoyancy ofthe Rumsey, the crew mayhave brokenup after dropping rising heat and force ofthe wind competed. The light theirtools at point 2. The officialreport states alltools fuels would have burned out quickly, leaving small were found within a 100-ft circle, so the crew was to- pockets ofbunchgrass or heavier stems to continue getheratthatpoint. Sallee saidthatNavon, aformer burning as the fire raced ahead. The high tempera- paratrooper who fought with the 82ndAirborne Divi- tures ofthe flames, 1,500 to 1,800 °F, would be lethal, sioninWorldWar II, couldbe seenup the slope ahead primarily because they would damage the respiratory ofthe crew. Others may already havejoined him, be- system. causethe final positionofthe bodies showedfourcrew Whenthe crew emergedfrom the trees intothe grass members traveled much farther thanthe others. In at point 3, Dodge must have realized they could not his book, Maclean describes these men as the "four reach safety and conceived the idea ofburning away horsemen." From this point, the analysis is divided a small clearing. This escape fire, as it has come to intotwo sections: the first describes the circumstances be called, wouldquickly clear an areawherethe crew ofthe crew members whose bodies were found close couldgo, afterthe fine fuels burned away, givingthem to Dodge, andthe seconddescribes the fourhorsemen. a chance to escape the flames ofthe main fire. From point 2, where the crew droppedtheirtools, to Dodge sized upthe situationbetterthanmost ofhis point3, whereDodge littheescapefire, is about240yd; crew, who either thought they could outrun the fire ifthe fire were spreading from 360 to 610 ft/min, it or saw no other alternative. Some ifnot all ofthe crew could have covered the distance in 1 to 2 min. Its stopped brieflyto see what Dodge was doing and lis- arrival time at point 3 would have been between 5:55 ten to his pleas forthem to get into the burned-out and 5:56. Ifthe slowest members died at about 5:56, area he was preparing. Someone is reported to have as indicatedby Harrison's broken watch, the estimated said: "To hell vdth that, I'm getting out ofhere!" No time ofthe fire's arrival is reasonable. We can setthe one stayed with Dodge. The crew members split up time thatthe mainfire overranDodge's escape fire at afterward, withthe majority continuingto runupthe about 5:55:30. canyon. I estimate they delayed no longer than 15 s This estimate is consistent with Dodge's statement at point 3, probably not that long. Some traveled on thatthe mainfirewasonly 50yd awaywhenhe stopped the contour and others went slightly downhill. The tolighthis escape fire. The mainfire could cover 50yd slowest ofthe crew members only got about 100 yd in 15 to 30 s ifit were spreadingfi-om 360 to610 ft/min. before being caught by the fire. One man broke his Ifthe men near Dodge died at 5:56, when Harrison's leg while fleeing on the steep, rocky slope. watch stopped, thenDodge andthose withhim arrived The fire couldhave covered 100 yd inlessthan a at point 3 where he lit his fire at about 5:55. minute at its calculated rate of600 to 750 ft/min. The crew took 2 min to go from point 2 where they Dodge estimatedthe menwere caughtin 30 s. Ifthey droppedtheirtools to point 3. Theycovered 240ydor had a 15-s lead on the fire after leaving him and trav- 720 ft, an average speedof360 ft/minor about 4 mi/h. eled 100 yd before being caught 45 s later, they would This is a very fast pace, considering the steep slope, havebeenrunningabout 400 fVmin or 4/4 mi/h, ahttle poorfooting, hightemperatures, andthe distance they fasterthan they were runningwhen they approached had already traveled. One man was reported to have Dodge at point 3. collapsed before reaching point 3. He had to be urged The four horsemen, who were found 375 yd beyond to continue. Dodge, may not have been with him when he started The official report describes the fuels at point 3 and his fire. We can examine how fast they would have beyond as cheatgrass and fescue. I assumed that the had to run to end up 375 yd beyond point 3, starting fuels were equally representedby fuel models 1 and 3. at either point 2 or point 3. From point 2 where the Model 1 is lightly loaded fuel that describes cheat- men dropped their tools to where the four horsemen grass very well; model 3 is a heavier grass model that diedis 620yd. Iftheywere at point 2 at 5:53, thefire, should accountforthefescue, weeds, andbrushinthe accordingto my estimate, would have caught them area. The fuel moisture was closeto 3 percent, avery at about 5:57,just a minute or so after it caught the 6