wildland fire behavior case studies and analyses: part 1 wildland fire ...

FireManagementtodayVolume 63 No. 3 Summer 2003

United States Department of AgricultureForest Service

WILDLAND FIREBEHAVIOR CASESTUDIES ANDANALYSES: PART 1

WILDLAND FIREBEHAVIOR CASESTUDIES ANDANALYSES: PART 1

Editors note: This issue of Fire Management Today reprints articles from early editions of the journal,some of them decades old. Although the articles appear in todays format, the text is reprinted largely ver-batim and therefore reflects the style and usage of the time. We made minor wording changes for clarity,added intertitles and metric conversions where needed, and occasionally broke up paragraphs or broke outsidebars to improve readability. All illustrations are taken from the original articles.

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 ofthe public 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 .

Ann M. Veneman, Secretary April J. BailyU.S. Department of Agriculture General Manager

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

Jerry Williams, Director Madelyn DillonFire and Aviation Management Editor

Carol LoSapioGuest Editor

Martin E. Alexander, Ph.D., and David A. Thomas Issue Coordinators

The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis ofrace, color, national origin, sex, religion, age, disability, political beliefs, sexual orientation, or marital or family sta-tus. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means forcommunication of program information (Braille, large print, audiotape, etc.) should contact USDAs TARGET Centerat (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 anequal opportunity provider and employer.

Disclaimer: The use of trade, firm, or corporation names in this publication is for the information and convenienceof the reader. Such use does not constitute an official endorsement of any product or service by the U.S. Departmentof Agriculture. Individual authors are responsible for the technical accuracy of the material presented in FireManagement Today.

ErratumIn the Spring 2003 issue of Fire Management Today, the table of contents showed an incorrect title forStephen J. Pynes article (volume 63[2], page 17). The correct title is Firestop II.

DEDICATION

This special issue of Fire Management Today is dedicated to the memory of Paul M. Gleason (19462003).

His passion for wildland firefighter safety and his deep professional interest in wildland fire behavior will be sorely missed.

Martin E. Alexander and David A. ThomasIssue Coordinators

http://www.fs.fed.us/fire/fmt/

CONTENTSWildland Fire Behavior Case Studies and Analyses: Value,

Approaches, and Practical Uses . . . . . . . . . . . . . . . . . . . . . 4M.E. Alexander and D.A. Thomas

Blackwater Fire on the Shoshone . . . . . . . . . . . . . . . . . . . . . . 9Division of Fire Control

The Factors and Circumstances That Led to the Blackwater Fire Tragedy . . . . . . . . . . . . . . . . . . . . . . . . . 11

A.A. Brown

Lessons From Larger Fires on National Forests, 1938 . . . . . . 15Roy Headley

Lessons From Larger Fires on National Forests, 1939 . . . . . . 23Roy Headley

Lessons of the McVey Fire, Black Hills National Forest . . . . . . 25A.A. Brown

An Analysis of the Honey Fire . . . . . . . . . . . . . . . . . . . . . . . 29C.F. Olsen

The Bower Cave Fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42Leon R. Thomas

The Possible Relation of Air Turbulence to Erratic Fire Behavior in the Southeast . . . . . . . . . . . . . . . . . . . . . . . . 46

George M. Byram and Ralph M. Nelson

The PinyonJuniper Fuel Type Can Really Burn . . . . . . . . . . . . 52Dwight A. Hester

A Firewhirl of Tornadic Violence . . . . . . . . . . . . . . . . . . . . . . 54Howard E. Graham

Rate of Spread on a Washington Fern Fire . . . . . . . . . . . . . . 56William G. Morris

Fire-Whirlwind Formation as Favored by Topography and Upper Winds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Howard E. Graham

Relationship of Weather Factors to Rate of Spread of the Robie Creek Fire . . . . . . . . . . . . . . . . . . . 63

R.T. Small

A Key to Blowup Conditions in the Southwest? . . . . . . . . . . . . 68Robert W. Bates

A Fire-Whirlwind in Alabama . . . . . . . . . . . . . . . . . . . . . . . . 71Gordon Powell

The Forest Fires of April 1963 in New Jersey Point the Way to Better Protection and Management . . . . . . . . . 74

Wayne G. Banks and Silas Little

The Harrogate FireMarch 15, 1964 . . . . . . . . . . . . . . . . . 79B.J. Graham

The Fire Behavior Team in Action: The Coyote Fire, 1964. . . . . 81John D. Dell

Gleason Complex Puts Up Huge Plume: A Tribute to Paul Gleason . . . . . . . . . . . . . . . . . . . . . . . . 85

Paul Keller

Interview With Paul Gleason . . . . . . . . . . . . . . . . . . . . . . . . 91Jim Cook and Angela Tom

SHORT FEATURESWebsites on Fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80Guidelines for Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 95

Volume 63 No. 3 Summer 20033

Firefighter and public safety isour first priority.

Volume 63 No. 3 Summer 2003Management todayFire

Historical photo from USDAForest Service files showingthe Wheeler Fire in 1948working down Bear Canyontoward Wheeler Gorge Campon the Los Padres NationalForest, CA. Photo: ForestService PhotographCollection, USDA ForestService, Washington Office,Washington, DC (no. 451594;F.E. Dunham, 1948).

The FIRE 21 symbol (shown below and on thecover) stands for the safe and effective use ofwildland fire, now and throughout the 21st century. Its shape represents the fire triangle(oxygen, heat, and fuel). The three outer red triangles represent the basic functions of wildland fire organizations (planning, operations,and aviation management), and the three criticalaspects of wildland fire management (prevention,suppression, and prescription). The black interiorrepresents land affected by fire; the emerginggreen points symbolize the growth, restoration,and sustainability associated with fire-adaptedecosystems. The flame represents fire itself as anever-present force in nature. For more informa-tion on FIRE 21 and the science, research, andinnovative thinking behind it, contact MikeApicello, National Interagency Fire Center, 208-387-5460.

On the Cover:

WILDLAND FIRE BEHAVIORCASE STUDIES AND ANALYSES: VALUE, APPROACHES, AND PRACTICAL USESM.E. Alexander and D.A. Thomas

Fire Management Today4

ince 1936, the WashingtonOffice of the USDA ForestService has published a period-

ical devoted to articles dealing witha very wide range of fire manage-ment topics. The name of this jour-nal has changed through the years,from Fire Control Notes, to FireManagement, to Fire ManagementNotes, and finally to Fire Manage-ment Today.* A good many of the243 issues that have been pub-lished in the past 67 years haveincluded a fire-behavior-relatedarticle. With the passage of time,however, many of these articleshave become buried, found onlyby the most intrepid researchers onthe shelves of major libraries.

In an effort to unbury the past andto increase both institutional mem-ory and organizational learningwithin the wildland fire communi-ty, the authors approached the edi-torial staff of Fire ManagementToday with the idea of republishinga selection of these past fire-behav-ior-related articles. We are pleasedthat they took us up on our sug-gestion.

Marty Alexander is a senior fire behaviorresearch officer with the Canadian ForestService at the Northern Forestry Centre,Edmonton, Alberta; and Dave Thomas isthe regional fuels specialist for the USDAForest Service, Intermountain Region,Ogden, UT.

* For more on the history of Fire Management Today,see Hutch Brown, How Did Fire Control NotesBecome Fire Management Today? Fire ManagementToday 60(1) [Winter 2000]: 814.

S In an effort to unbury the past and to increase both institutional memory and organizational learning within the wildland fire community, we are reprinting past articles

on fire behavior.

This special issue of Fire Manage-ment Today begins a series of threeconsecutive issues with articlesrelated to fire behavior. This issuecontains the first of two install-ments of articles involving firebehavior case studies and analysesof wildfires; examples pertaining toprescribed fires are not included(e.g., Custer and Thorsen 1996).The 19 case studies and analyses inthis issue are presented in chrono-logical order, from 1937 to 1967.The third issue in this series willbe devoted to aids, guides, andknowledge-based protocolsinvolved in forecasting wildlandfire behavior for safe and effectivefire suppression.

General Value of CaseStudies The importance of documentedcase studies or histories of wildlandfires has been repeatedly empha-sized by both fire managers andfire researchers (e.g., Byram 1960;Thomas 1994; Turner and others1961). As long-time Forest Servicewildland fire researcher/adminis-trator Craig Chandler (1976) hasnoted, Time and time again casehistories have proven their value astraining aids and as sources ofresearch data. The authors strong-

ly support this notion and haveendeavored to reflect it in our indi-vidual work areas in fire researchand fire management, respectively(Alexander and Lanoville 1987;Thomas 1991).

The idea of relying on wildfires as apossible source of data is especiallypertinent to empirically basedschemes for quantitative firebehavior prediction that rely onthis kind of information in wholeor in part (e.g., Alexander 1985;Forestry Canada Fire DangerGroup 1992; Rothermel 1991).This fact is especially significant atthe extreme end of the fire intensi-ty scale, where experimental firesare exceedingly difficult to arrange(Alexander and Quintilio 1990;Cheney and others 1998).

Information gleaned from wildlandfire behavior case studies has alsoproved of value in testing and eval-uating various fire models, theo-ries, decision aids and support sys-tems, and management guidelines(e.g., Anderson 1983; Haines andothers 1986; Nelson 1993; Pearceand Alexander 1994). For example,Lindenmuth and Davis (1973) usedan observation of the initial run ofthe Battle Fire, a 28,400-acre


(11,500-ha) fire that occurred May1420, 1972, on the PrescottNational Forest, AZ, to assess theperformance of their empiricallybased model for predicting firespread in Arizonas oak chaparralfuel type.

Approaches to CaseStudiesThere are many examples in NorthAmerica and elsewhere where fireresearchers and fire managers haveattempted to observe and docu-ment the behavior of free-burningfires, using various types of datacollection methods and monitor-ing equipment, on an ad hoc or amore formal basis (e.g., Barneyand others 1978; Barrows 1961;Billing 1986; Schaefer 1957;Traylor 1961*). These effortsextend back many years (Gisborne1929) and continue into the 21stcentury (e.g., Burrows and others2000).

Fire behavior researchers arerarely in the right place at theright time to observe and docu-ment the behavior of forest andrange fires. While there have, ofcourse, been some exceptions (e.g.,Sneeuwjagt 1974; Stocks andFlannigan 1987), including escapesfrom outdoor experimental fires(Alexander and others 1991; Stocks1987), for the most part fire opera-tions personnel tend to be in thebest position to make and recordkey observations. Probably themost concerted and systematiceffort by fire researchers to observeand record actual fire behavior wasmade by the Forest ServicesSouthern Forest Fire Laboratory inMacon, GA, from the late 1950s toearly 1970s (DeCoste and Sackett

* A summary of this work can be found in R.E. Traylor,Correlation of Weather to Fire Spread in Grass andBrush Fuels on the Snake River Plains in SouthernIdaho, Fire Control Notes 22(4) [Fall 1961]: 118119.

1966; Sackett and DeCoste 1967).This was no doubt due in large partto George Byrams (1960) influ-ence.

Some limited documentation hasalso been undertaken by fire man-agers and fire researchers servingas fire behavior officers or special-ists/analysts on various wildlandfire incidents (e.g., Johnson 1964;McCaw, Maher, and Gillen 1992;Norum 1982; Thomas 1991). Fireresearchers have also been involvedin many after-the-fact investiga-tions (e.g., Butler and Reynolds1997; Fogarty and others 1996;McCaw, Simpson, and Maher1992). Van Wagner (1971) haspointed out that some valuablereference data can be collected bybeing in the right place at the righttime. It is, in fact, quite feasible toobtain good data by visiting thescene of a fire shortly after ithas occurred, while its history isstill fresh both on the ground andin the mind of the fire boss.

Byram (1954) made extensive useof the case study method of indi-vidual fires in his research intoblowup fire behavior. As he notes,Some of the observations anddetails of behavior are writtendown in fire reports, but most ofthe information is still in thememories of men who worked onthe fires. Fire behavior may, there-fore, be difficult to reconstruct attimes, especially on fires whichoccurred a number of years ago.Usually, however, a surprisingamount of detail can be obtainedby talking with men who were onthe fires and by going over the firearea with them.

A final possibility is the hindsightanalysis of major wildland fire inci-dents in the light of present-dayknowledge and tools using existing

historical information to establishthe fires chronology and generalbehavior. The reports of Haines andSando (1969), Stocks and Walker(1973), Street and Alexander(1980), and Rothermel (1993) aregood examples of this approach tocase studies.

Pragmatic Value ofCase StudiesA practical fire manager, alwaysinterested in the control of wild-fires and the ignition of prescribedfires, might ask: What is the use ofhistorical fire behavior case stud-ies? How can old documents helpfire management personnelbecome better managers of forestand range fires, in all their forms?Beyond the recreation of a goodread, what utility do these articlesoffer? How can old essays becomerelevant for a 21st-century fire-fighter?

The old articles will only seemdated if we fail to make use ofthem. There are two primary rea-sons to thoroughly study these firebehavior case studies:

To learn from them and therebylessen the chance of making thesame mistake again; and

To prepare ourselves not to besurprised to the point of distrac-tion by a fires surprising behav-ior in a particular fuel type undera given weather condition.

Not making the same mistaketwice and being prepared to be sur-prised will go a long way towardcreating a highly reliable firefight-ing organization where safety trulymatters.

Unless we actively learn from pastwildland fires, then the only waywe can gain additional fire behav-ior knowledge is to actually experi-


ence a fires behavior or to modelthe fires behavior on a computerat our desk. Even the most activefire behavior analyst (FBAN) rarelygets enough near-real-time oppor-tunities to predict the spread andintensity in every fuel complex orto complete a prediction enoughtimes to become good at it(Thomas 1994). The best learningscenario for a practicing fire behav-iorist is a combination of all threelearning techniques: actively usingcase studies, getting field experi-ence, and practicing computermodeling. Each is a distinct modeof learning and adaptation; whencombined, they become a powerfulmodel for continuous learning.

Case study knowledge, coupledwith experienced judgment and firebehavior modeling, is also consid-ered an effective operational tech-nique or procedure for appraisingfire potential (Brown 1978).Burrows (1984) maintains thatmost wildland firefighters basetheir expectations of how a fire willbehave largely on experience and,to a lesser extent, on fire behaviorguides. If this is indeed the case,then it is worth reiterating thepoints made by Forest Service fireresearch pioneer Harry T. Gisborne(1948) about experienced judg-ment: For what is experiencedjudgment except opinion based onknowledge acquired by experience?If you have fought forest fires inevery different fuel type, under allpossible kinds of weather, and ifyou have remembered exactly whathappened in each of these combi-nations, your experienced judg-ment is probably very good. But ifyou have not fought all sizes offires in all kinds of fuel types underall kinds of weather then your

experience does not include know-ledge of all the conditions.Presumably then, case studies canhelp supplement and therebystrengthen (but never replace) apersons experience level.

Safety Value of CaseStudiesAs we read through this chronolog-ical selection of articles, especiallythe accounts of forest fires wherefirefighters lost their lives or therewere near-misses or unforeseenblowups, we can ask ourselves andour crews whether we have fullygrasped the major lessonslearned from these past firebehavior events. Excellent methodsof using past fire behavior knowl-edge from case studies to increasewildland firefighter safety in thefuture are the staff ride (Alexander2002; Thomas and Cook 2002),*the sand box exercise (Euler 1946),yearly fire refreshers (e.g., the 2001USDI Bureau of Land ManagementFireline Safety Refresher videos),and weekly tailgate safety meet-ings.

For example, one of these articlescould be handed out each week tomembers of an organized fire crew.The crew would be given time toread and ponder the article. Then,in a group setting, with the fireforeman (i.e., hotshot superintend-ent, smokejumper-in-charge, localfire management officer, etc.) act-ing as facilitator, the crew could beled through a series of questionsthat the article has inspired. Forexample:

* For more on the staff ride technique, see the variousarticles on the Dude Fire Staff Ride in FireManagement Today 62(4) [Fall 2001].

Is there something that we canapply to our current situation?

Have we learned all that this oldfire has to teach us?

Could the same situation occurtoday?

What are we going to do differ-ently after reading this casestudy?

This process, if faithfully followedthroughout a fire season, wouldincrease both mindfulness andresilience (Weick and Sutcliffe2001), the two hallmarks of indi-viduals and their organizationsdetermined to do everything theycan to control and use wildlandfire safely.

Both authors have used case stud-ies to lead training sessions in theclassroom. One of us (Thomas) hasalso used the technique in the fieldat the site of past fires. In June1994, a group of FBANs on a visitto the site of the 1949 Mann GulchFire were asked, using existing his-torical case study information as astarting point for a fire behaviorprediction, if they could have pre-vented the firefighter fatalities thatoccurred on this infamous fire.Using the available historical fireinformation, a similar questionwas asked of a large group of firemanagement personnel on a staffride of the 1990 Dude Fire(Thomas and Cook 2002). In bothof these examples, many of the stu-dents said that these trainingsessions were some of the best theyhad ever attended. Using case stud-ies or histories, an old fires firebehavior came alive.

Time and time again case histories have proven their value as training aids

and as sources of research data. Craig Chandler (1976)


Another benefit of having thesearticles available again is for theiruse within fuel specialist reportsused in environmental assess-ments. Fuel specialists are increas-ingly called upon to justify why aninterdisciplinary team recom-mended one fire hazard abatementtechnique over another. These casehistories, especially the descrip-tions of fire behavior in a givenfuel type (e.g., Helms 1979), couldbe cited in those reports (or hyper-linked to a main database), savingmuch analysis time. The fuels spe-cialist would not have to explainhow a fire might burn in a givenfuel type, for she or he would havea published account to cite orhyperlink to.

Learning ContributionA learning organization has beendefined as one that is skilled atcreating, acquiring, interpreting,transferring, and retaining knowl-edge, and at purposefully modify-ing its behavior to reflect newknowledge insights (Garvin 2000).Fire behavior case studies go along way toward preparing a foun-dation for organizational learning;in so doing, they follow the truespirit of learning implied in thisdefinition. Simply put, our firemanagement culture, now domi-nated by a learning pattern of trialand error, would become a learn-ing culture, one in which a sys-tematic study of the past throughthe use of case studies wouldbecome a routine procedure.

This special issue of Fire Manage-ment Today devoted to fire behav-ior, and the two others that will

follow, are in keeping with theideals and sentiment expressed byRoy Headley (1936) in the very firstissue of Fire Control Notes.Headley, who cofounded the jour-nal as the head of the ForestServices Division of Fire Control(the predecessor of todays Fire andAviation Management), called forintegrating and sharing the expe-rience, thinking, and experimentsof the many people engaged inwildland fire management. To thisend, Headley envisioned FireControl Notes as a common meet-ing ground, a clearing-house ofdevelopments. In this sense, FireManagement Today, by republish-ing the past (and thereby revivingit for the future), has rediscoveredits own unique niche.

AcknowledgmentsThe authors offer their sincerestheartfelt appreciation to HutchBrown, Madelyn Dillon, and CarolLoSapio, editors of Fire Manage-ment Today, for their significantcontributions to this special issue,and to April Baily, the journalsgeneral manager, for supportingthe concept of these special issueson wildland fire behavior. Theirdedication and outstanding editori-al abilities have brought life tomany of the articles contained inthis issue that have long been forgotten.

References Alexander, M.E. 1985. Estimating the

length-to-breadth ratio of elliptical forestfire patterns. In: Donoghue, L.R.; Martin,R.E., eds. Proceedings of the EighthConference on Fire and ForestMeteorology; 1985 April 29May 2;Detroit, MI. SAF Publ. 8504. Bethesda,

MD: Society of American Foresters:287304.

Alexander, M.E. 2002. The staff rideapproach to wildland fire behavior andfirefighter safety awareness training. FireManagement Today. 62(4): 2530.

Alexander, M.E.; Lanoville, R.A. 1987.Wildfires as a source of fire behaviordata: A case study from NorthwestTerritories, Canada. In: PostprintVolume, Ninth Conference on Fire andForest Meteorology; 1987 April 2124;San Diego, CA. Boston, MA: AmericanMeteorological Society: 8693.

Alexander, M.E.; Quintilio, D. 1990.Perspectives on experimental fires inCanadian forestry research. Mathematicaland Computer Modelling. 13(12): 1726.

Alexander, M.E.; Stocks, B.J.; Lawson, B.D.1991. Fire behavior in blacksprucelichen woodland: The Porter LakeProject. Inf. Rep. NORX310.Edmonton, AB: Forestry Canada,Northern Forestry Centre.

Anderson, H.E. 1983. Predicting wind-driv-en wild land fire size and shape. Res. Pap.INT305. Ogden, UT: USDA ForestService, Intermountain Forest and RangeExperiment Station.

Barney, R.J.; Noste, N.V.; Wilson, R.A. 1978.Rates of spread of wildfire in Alaskanfuels. Res. Note PNW311. Portland, OR:USDA Forest Service, Pacific NorthwestForest and Range Experiment Station.

Barrows, J.S. 1961. Natural phenomenaexhibited by forest fires. In: Berl, W.G.,ed. Proceedings of the InternationalSymposium on the Use of Models in FireResearch; 1960 November 910;Washington, DC. Publ. 786. Washington,DC: National Academy of SciencesNational Research Council.

Billing, P. 1986. Operational aspects of theinfra-red line scanner. Res. Rep. No. 26.Melbourne, VIC: Victoria Department ofConservation, Forests & Lands, FireProtection Branch.

Brown, J.K. 1978. Fuel inventory andappraisal. Paper presented at the USDAForest Service National Fire-Danger andFire-Weather Seminar; 1972 November1416; Missoula, MT.

Burrows, N.D. 1984. Predicting blow-upfires in the jarrah forest. Tech. Pap. No.12. Perth, WA: Forests Department ofWestern Australia.

Burrows, N.; Ward, B.; Robinson, A. 2000.Behavior and some impacts of a largewildfire in the Gnangara maritime pine(Pinus pinaster) plantation, WesternAustralia. CALMScience. 3: 251260.

Butler, B.W.; Reynolds, T.D. 1997. Wildfirecase study: Butte City Fire, southeasternIdaho, July 1, 1994. Gen. Tech. Rep.INTGTR351. Ogden, UT: USDA ForestService, Intermountain Research Station.

A surprising amount of detail can be obtained by talking with men who were on the fires and by going over the fire area with them.

George Byram (1954)


Byram, G.M. 1954. Atmospheric conditionsrelated to blowup fires. Stn. Pap. No. 35.Asheville, NC: USDA Forest Service,Southeastern Forest Experiment Station.[Reprinted as: National Fire EquipmentSystem Publication NFES 2565 by theNational Wildfire Coordinating Group,Boise, ID.]

Byram, G.M. 1960. A problem analysis andproposed research program for theSouthern Forest Fire Laboratory. Macon,GA: USDA Forest Service, SoutheasternForest Experiment Station, SouthernForest Fire Laboratory.

Chandler, C.C. 1976. Meteorological needsof fire danger and fire behavior. In:Baker, D.H.; Fosberg, M.A., tech. coords.Proceedings of the Fourth NationalConference on Fire and ForestMeteorology; 1976 November 1618; St.Louis, MO. Gen. Tech. Rep. RM32. FortCollins, CO: USDA Forest Service, RockyMountain Forest and Range ExperimentStation: 3841.

Cheney, N.P.; Gould, J.S.; Catchpole, W.R.1998. Prediction of fire spread in grass-lands. International Journal of WildlandFire. 8: 113.

Custer, G.; Thorsen, J. 1996. Stand-replace-ment burn in the Ocala NationalForestA success. Fire ManagementNotes. 56(2): 712.

DeCoste, J.H.; Sackett, S.S. 1966. Baptismby fire. Southern Lumberman.213(2656): 169170.

Euler, D.H. 1946. The sand box as a fire-control training tool. Fire Control Notes.7(1): 3739.

Fogarty, L.G.; Jackson, A.F.; Lindsay, W.T.1996. Fire behaviour, suppression andlessons from the Berwick Forest Fire of26 February 1995. FRI Bull. No. 197, For.Rural Fire Sci. Tech. Ser. Rep. No. 3.Rotorua and Wellington, NZ: NewZealand Forest Research Institute andNational Rural Fire Authority.

Forestry Canada Fire Danger Group. 1992.Development and structure of theCanadian Forest Fire Behavior PredictionSystem. Inf. Rep. STX3. Ottawa, ON:Forestry Canada, Science andSustainable Development Directorate.

Garvin, D.A. 2000. Learning in action: Aguide to putting the learning organiza-tion to work. Boston, MA: HarvardBusiness School Press.

Gisborne, H.T. 1929. The complicated con-trols of fire behavior. Journal of Forestry.27: 311312.

Gisborne, H.T. 1948. Fundamentals of firebehavior. Fire Control Notes. 9(1): 1324.

Haines, D.A.; Main, W.A.; Simard, A.J. 1986.Fire-danger rating and observed wildfirebehavior in the northeastern UnitedStates. Res. Pap. NC274. St. Paul, MN:USDA Forest Service, North CentralForest Experiment Station.

Haines, D.A.; Sando, R.W. 1969. Climaticconditions preceding historically greatfires in the North Central Region. Res.Pap. NC84. St. Paul, MN: USDA ForestService, North Central ForestExperiment Station.

Headley, R. 1936. Fire Control Notes offersits services. Fire Control Notes. 1(1): 34[reprint: Fire Management Today. 60(1):67].

Helms, J.A. 1979. Positive effects of pre-scribed burning on wildfire intensities.Fire Management Notes. 40(3): 1013.

Johnson, V.J. 1964. Chronology and analy-sis of the Hughes Fire, 1962. Res. NoteNOR8. Juneau, AK: USDA ForestService, Northern Forest ExperimentStation.

Lindenmuth, A.W., Jr.; Davis, J.R. 1973.Predicting fire spread in Arizonas oakchaparral. Res. Pap. RM101. FortCollins, CO: USDA Forest Service, RockyMountain Forest and Range ExperimentStation.

McCaw, L.; Maher, T.; Gillen, K. 1992.Wildfires in the Fitzgerald River NationalPark, Western Australia, December 1989.Tech. Rep. No. 26. Perth, WA:Department of Conservation and LandManagement.

McCaw, L.; Simpson, G.; Mair, G. 1992.Extreme wildfire behaviour in 3-year-oldfuels in a Western Australian mixed euca-lyptus forest. Australian Forestry. 55:107117.

Nelson, R.M., Jr. 1993. Byrams derivationof the energy criterion for forest andwildland fires. International Journal ofWildland Fire. 3: 131138.

Norum, R.A. 1982. Predicting wildfirebehavior in black spruce forests ofAlaska. Res. Note PNW401. Portland,OR: USDA Forest Service, PacificNorthwest Forest and Range ExperimentStation.

Pearce, H.G.; Alexander, M.E. 1994. Firedanger ratings associated with NewZealands major pine plantation wildfires.In: Proceedings of the 12th Conferenceon Fire and Forest Meteorology; 1993October 2628; Jekyll Island, GA. SAFPubl. 9402. Bethesda, MD: Society ofAmerican Foresters: 534543.

Rothermel, R.C. 1991. Predicting behaviorand size of crown fires in the NorthernRocky Mountains. Res. Pap. INT438.Ogden, UT: USDA Forest Service,Intermountain Research Station.

Rothermel, R.C. 1993. Mann Gulch Fire: Arace that couldnt be won. Gen. Tech.Rep. INT299. Ogden, UT: USDA ForestService, Intermountain Research Station.

Sackett, S.S.; DeCoste, J.H. 1967. A newmobile fire laboratory. Fire ControlNotes. 28(4): 79.

Schaefer, V.J. 1957. The relationship of jetstreams to forest wildfires. Journal ofForestry. 55: 419425.

Sneeuwjagt, R.J. 1974. Evaluation of thegrass fuel model of the National FireDanger Rating System. M.S. thesis.Seattle, WA: University of Washington.

Stocks, B.J. 1987. Fire potential in thespruce budworm-damaged forests ofOntario. Forestry Chronicle. 63: 814.

Stocks, B.J.; Flannigan, M.D. 1987.Analysis of the behavior and associatedweather for a 1986 northwesternOntario wildfire: Red Lake No. 7. In:Postprint Volume, Ninth Conference onFire and Forest Meteorology; 1987 April2124; San Diego, CA. Boston, MA:American Meteorological Society:94100.

Stocks, B.J.; Walker, J.D. 1973. Climaticconditions before and during four signif-icant forest fire situations in Ontario.Inf. Rep. OX187. Sault Ste. Marie, ON:Canadian Forestry Service, Great LakesForest Research Centre.

Street, R.B.; Alexander, M.E. 1980.Synoptic weather associated with fivemajor forest fires in Pukaskwa NationalPark. Int. Rep. SSD802. Toronto, ON:Environment Canada, AtmosphericEnvironment Service, Ontario Region.

Thomas, D.A. 1991. The Old Faithful Innfire run of September 7, 1988. In:Andrews, P.L.; Potts, D.F., eds.Proceedings of the 11th Conference onFire and Forest Meteorology; 1991 April1619; Missoula, MT. SAF Publ. 9104.Bethesda, MD: Society of AmericanForesters: 272280.

Thomas, D. 1994. A case for fire behaviorcase studies. Wildfire. 3(3): 45, 47.

Thomas, D.; Cook, W. 2002. Dude Firestaff ride. Fire Management Today. 62(4):45.

Traylor, R.E. 1961. Correlation of weatherto fire spread in grass and brushlandfuel types on the Snake River Plains ofsouthern Idaho. M.S. thesis. Missoula,MT: Montana State University.

Turner, J.A.; Lillywhite, J.W.; Pieslak, Z.1961. Forecasting for forest fire services.Tech. Note No. 42. Geneva, Switzerland:World Meteorological Organization.

Van Wagner, C.E. 1971. Two solitudes inforest fire research. Inf. Rep. PSX29.Chalk River, ON. Canadian ForestryService, Petawawa Forest ExperimentStation.

Weick, K.E.; Sutcliffe, K.M. 2001.Managing the unexpected: Assuring highperformance in an age of complexity.San Francisco, CA: Jossey-BassA WileyCompany.


BLACKWATER FIRE ON THE SHOSHONE*USDA Forest Service, Division of Fire Control

reliminary reports in hand asthis issue goes to press showthat initial action on this light-

ning fire was alert, prompt and vig-orousquite remarkably so, con-sidering that the ShoshoneNational Forest is rated as a low-danger forest, and doesnt evenhave lookout stations. The countrywas high and steepjust belowtimber line. In spots the lodgepolepine and fir were dense andlimbythe familiar patches ofrather scrubby jungle found on thebetter sites at high elevations.There were steep slopes coveredwith dense but not jungly standsjust the setting for wind-drivencrown fires of intense heat.Pictures of the area show bareridge tops and open places hereand there. Fuel on the groundseems to have been quite lightaswould be normal under such con-ditions. One would guess that thefuel experts would rate the area atLow rate of spread and Lowresistance to control. But whenthe heavy wind started sweepingthis way and that on Saturday,August 21, fifteen men lost theirlives. Six of these died from theirburns after the blow-up.

Large Loss of LifeThe danger from such accidentsprobably is statistically less thanthe danger from automobile acci-dents, which is so familiar we

When the heavy wind started sweeping this way and that on Saturday, August 21,

fifteen men lost their lives.

* The article is reprinted from Fire Control Notes 1(5)[20 September 1937]: 305306.

The Division of Fire Control in the USDAForest Services national office was theprecursor of the agencys Fire and AviationManagement Staff, Washington Office,Washington, DC.

Plargely ignore it. But such fireaccidents do happen and impressus all the more because of theirinfrequency. This is the largest lossof life from a single National Forestfire since 1910. It is the irony offate that it had to occur on aNational Forest which, so far ascan be determined from therecords here, has had only oneother large fire during its wholehistory. The latest reports on sizeof this fire put it at 1,100 acres(445 ha).

To the men who died in this disas-ter, all fire control men everywherepay tribute. To the bereaved fami-lies they extend the deepest sympa-thy. To the survivors, and particu-larly the exceptionally large num-ber of injured men, is extendedappreciation and cordial concernfrom all those engaged in the highadventure of protecting Americanforests from devastation by fire.

District Ranger Posts statement(excerpted in the sidebar on page10) is published as an authenticcase record of the processes ofjudgment in such situations wherea man must think first and thinkclearly about the safety of the menin his crew. His words will recall toall experienced men many days ofharassed effort to get the lineahead and the fire mopped upbefore something happenedbut

always with a running accompani-ment of a plan (sometimes uncon-scious) for the best way to safetyfor the crew if something wentwrong.

His statement is also a technicalcase history of the handling of menin such crises. His record could befollowed better with the aid of amap, but even without it much canbe gleaned from the story.

Heroic ConductAs a record of unassuming heroicconduct the statement needs nocomment. It was dictated straight-away in the presence of D.P.Godwin, with no rehashing or edit-ing except the correction of thespelling of one name and the inser-tion of the name of Bert Sullivanin one place. Post has some badburns, and both hands and bothsides of his face are heavily band-aged. He is out of danger, but willbear scars.

Junior Forester Tyrrell of RangerPosts party died later from hisburns. In speaking of him in hissigned statement, Enrollee AlcarioSerros says:

Then we saw that we didnt haveno chance to go back, so RangerPost told Mr. Tyrrell to take care ofus, and he took us up to the rimrock. The fire started from the


east, and then south,and then the west. Itwas the west fire thatburned us. As the firecame closer to us welayed down on the rockridge. Mr. Tyrrell layedon top of me. When thefire burned Mr. Tyrrellhe ran and I ran, too,about 10 feet.

District RangerClayton, whose mes-sage Ranger Postreceived, died with sixof his men. A seventhgot out, but died fromhis burns.

When cut off from the safety sought above timber line (center background), Ranger Post and 40 menwho followed his instructions survived by taking what shelter they could, first on the left (north) side,then the right side of the bare spot on the ridge (center foreground). Junior Forester Tyrrell died laterfrom burns inflicted by the flames and heat that swept over them.

The Blowup Begins

The following eyewitness account* describes the onset of the blowup on the Blackwater Fire on the after-noon of August 21, 1937. The blowup resulted from the passage of a dry cold front.

Up until this time no wind was inevidence. Almost like a shot outof a gun, there was a heavy wind.It swept through the area in asnear as I can determine in anortheasterly direction, this car-ried sparks over the constructedline and below us. I heard a fireroar to the northwest and itappeared to be a considerable dis-tance away. I called to Tyrrell andtold him that something wasgoing wrong and that I was goingto investigate.

I ran for some distance to thenorthwest and climbed a rockypoint and saw below me a spot fireof considerable size burning to thenortheast and around a ridge tothe north of us. My impression wasthat this fire should be immediate-ly taken care of and possibly aban-don work on our line in order todo it.

I turned to summon help for thisjob when there was decided changein the wind again and the spot firewas swept into the southwest

directly into the line of men online construction. In a few sec-onds numerous fires appearedbelow the line.... Almost at onceit was clearly evident that furtherattempts at line construction inthat area was out of the question.I sent out a call for all men toabandon their work and proceedto the ridge to the northeast.This was approximately threeoclock, P.M.

U.J. Post, District Forest RangerAugust 27, 1937

* Excerpt from Statement by Ranger Urban J.Post, Fire Control Notes 1(5) [20 September 1937]:308315. Photos taken of the blowup can be foundin E. Kauffman, Death in Blackwater Canyon,American Forests 43(11) [November 1937]:534540.


THE FACTORS AND CIRCUMSTANCES THATLED TO THE BLACKWATER FIRE TRAGEDY*A.A. Brown

Four Critical Factorshile no proof is available,the nature and circum-stances of the blow-up on

August 21 seem to indicate that anundiscovered spot fire, probablyfrom the night before, to the northof Clayton Gulch and over thesmall, sharp ridge in Clayton Creek(one-half mile [0.8 km] west andnorth of the point where theClayton group was later found),was the first critical factor in mak-ing the trap in which the men werecaught and burned to death onthat day. Apparently this spot fireat first spread up the slope imme-diately above to the northeast. Thisis clearly indicated by the noteRanger Clayton sent to RangerPost at the time the spot firespread conspicuously just prior tothe blow-up.

The second critical circumstancewas the fact that the timber abovethe newly constructed line had notcrowned out except for a smallfringe along the south edge. Thethird critical circumstance was thefact that spotting from the fire ofthe previous day had given a

The nature and circumstances of the blow-up on August 21 indicate

that there were four critical factors involved.

* The article is reprinted from Fire Control Notes 1(6)[December 6, 1937]: 384387.

Original editors note: Included as a vital part of the full report on the Blackwater Fire was the report made byA.A. Brown after an exhaustive study of the fire behavior and the critical circumstances which converged tobring about the tragedy. Mr. Brown had been only recently transferred to Region 2 to head the fire controlwork of that Region, and he brought to the tasks involved in this disaster the sound knowledge and discern-ment springing from long and successful fire experience.

When this article was originally published,A.A. Brown was head of Fire Control forthe USDA Forest Service, Rocky MountainRegion.

ragged edge to the burning area onthe steep downhill side, with smallspots below the general front. As aresult the fire fighters found itexpedient to connect the fire linebelow the hottest spots, leavingconsiderable unburned surface fuelinside the line at the lowest point.

The fourth critical element was thenature of the forest fuel in thisdrainage. It consisted of a verydense stagnated stand of Douglasfir with a varying mixture5 to 15per centof spruce and of alpinefir. A dense overhead canopy exist-ed, with dead branches nearly tothe ground, with many small,brushy, dead or nearly dead sup-pressed trees as an understory, aconsiderable volume of sound deadbranches, logs and suppressedtrees on the ground, and with vary-ing amounts of moss throughoutthe canopy and on all the deadbranches. In addition, slopes of 20 to 60 per cent prevailed.

These four factors set the stage forwhat happened.

Strong WindThe relative humidity at 1 p.m. onAugust 21 was 6 per cent, with atemperature of 90 F (32 C) at thefire danger station at the WapitiCivilian Conservation Corps (CCC)camp, two and one-half miles (4km) away at 2,000 feet (610 m)lower elevation. At approximately3:30 p.m., with these critical cir-cumstances prevailing, a stronggusty wind of apparently at least 30 miles velocity per hour (48 km/h) came up from the south-west. About 3:45 p.m. it swervedand became a west wind. (Thesetimes are based on the circumstan-tial evidence of other events of thefire.)

The duration of this strong velocityis uncertain because of the strongconvectional winds set up almostat once by the crowning. It is rea-sonable to suppose that the changein direction may have been largelya convectional effect. At the start,timber began to crown above theline and the whole fire there beganto pick up in intensity and tothrow new spots below the line, asmight be expected. Possibly this

W


The two crown fires then rapidly closed togetherwith the cyclonic effect of such a circumstance,sweeping clean the major portion of the head

of the Clayton Creek drainage.

exerted a strong convectional pullon the spot fire below, which hadalso begun to crown.

At any rate, the course of the drivefrom the spot fire changed to theeast and started directly up thedrainage. The two crown fires thenrapidly closed together with thecyclonic effect of such a circum-stance, which reached its climax at4:20 p.m. As a result, the majorportion of the head of the ClaytonCreek drainage from the spot fireup to Double Mountain was sweptclean in a final crown fire confla-gration which was completed byapproximately 5 p.m.

In this conflagration 9 deathsoccurred directly. Six additionalmen were so badly burned thatdeath ensued, and 36 additionalmen suffered injuries from whichthey are recovering.

Firefighter Movements*Just before the crowning started,the distribution of men on thenewly constructed line in ClaytonCreek had been as follows: Fivemen of the National Park Servicecrew on mop-up were operating asfar as the first small park, about 5chains (330 feet [101 m]) northeastof the ridge. Beyond them for 30chains (1,980 feet [603 m]) were 6men of the BPR crew, who hadbeen actively pushing the new lineconstruction from this point on,and who had got as far as ClaytonGulch, plus a few men who hadbeen dropped off from Posts crew.Beyond them were Ranger Post,with Foremen Tyrrell and Sabanand sub-foreman Hale, with about40 men who had taken up the newline at Clayton Gulch and hadcompleted 16 chains (1,056 feet

[322 m]) at the time the blow-upoccurred.

Clayton, who had been placed incharge as sector boss of the newconstruction, was following thefresh crew in and checking up onconditions as he went. Apparentlyhe was checking particularly onspot fires. The BPR crew were giv-ing most of their attention to spotfires at two points below the 30chains (1,980 feet [603 m]) of linethey had constructed. They wereabout 20 chains (1,320 feet [402m]) in from Trail Ridge, except forPierce, one of their members whohad been left alone on hot line at apoint about 10 chains (660 feet[201 m]) in, where several logswere on fire close to the firetrench. Two men were left to helphim as Posts crew came past andabout 6 men were left with Sabanand Clayton to work on spot fires.

By the time this distribution wascompleted, about 3:45 p.m., Post,Clayton, and Fifield, probablysimultaneously, saw evidence of anuncontrolled spot fire. Fifield,according to his statement, was onthe rock point of Trail Ridge at thetime and thought at first that itwas the spot near the bottom ofthe first gulch which had previous-ly been found and trenched, butdiscovered instead that it was inline with it, but over the smallridge just to the north. He at oncegave thought to Wolcotts crew,who were in this vicinity, butfound them coming out onaccount of the crowning there.Wolcott immediately went on up

Trail Ridge and also called out themen from the fire trail north ofTrail Ridge. Pierce, who had beennear the highest point of the firetrail before it dipped down intoClayton Gulch, had already comeout to the first small park with thetwo CCC boys helping him becauseof a flash of crowning just belowhim, which apparently crossed thefire line but died down again at thelittle park. He attempted to get theattention of the rest of the BPRcrew, but, receiving no answer,decided they were withdrawing theother way.

About 6 other CCC boys were alsoassembled at the park, and all cameout together at Wolcotts alarmcall. The heavy crowning apparent-ly occurred shortly afterward(about 4 p.m.). Posts attention wasattracted to the spot fire when itstarted crowning toward the north-east up the slope on the north sideof Clayton Creek. His first thoughtwas to take his crew to it, but thewind changed and the fire startedup the gulch before he could takeany action to that end. Accordinglyhe started moving his crew fromits path as best he could, asdescribed in his statement.

Death TrapClaytons movements are not soclear in detail. It is evident fromthe note he dispatched to Post thatthe spot fire had attracted hisattention, apparently from on thespur ridge just south of the gulch,where he was later trapped. Up tothe time this fire started directlyup the gulch it was a threat to the*See the maps on page 14 (from the article by

D.P. Godwin referenced in the sidebar on page 13).


With the direction of the fires path directly up the gulch, it probably acted as a

furnace draft and became a death trap.

line above which [it] must bestopped, but probably did notappear to be dangerous to life.Clayton saw it was the focal point ifthe line were to be held, and thatmore men than the 7 with himwould be required.

It does not seem likely that hewaited on the ridge above ClaytonGulch the 20 minutes or more thatseems to have elapsed from thefirst active spreading of the spotfire below until the general blow-up occurred. Probably he starteddown toward it, either with hisgroup or alone. If alone, he proba-bly left instructions for his groupto await his investigation of poten-tialities below. Or, if he took hisgroup, he probably left one ormore men at the spring in thegulch to await reinforcements fromPost. In either case, the naturalroute of travel toward the spot firewould be down the gulch toward it.

Once off the ridge the full poten-tialities of the fire below would nothave been immediately as evidentto him as it was to Post above.Presumably, as soon as he sawwhat was about to happen, heturned back to get the men at thespring. In doing so he had to goback up the slope ahead of the fireand probably got to his men just atthe time that every avenue ofescape was cut off. Had the firebeen going across topography thebare gulch might have served satis-factorily as a refuge. With thedirection of its path directly up thegulch, it probably acted as a fur-nace draft and became a death trap.

Reconstructed TragedyIn conclusion, the reconstructedtragedy depended on each of thefour factors first discussed whichcontributed to the behavior of thefire, plus the distribution andmovement of the men at the time.The high wind and burning condi-

tions alone, without the spot fire,would have created a dangeroussituation, and would have no doubtforced abandonment of the newlyconstructed fire line, but withoutloss of life, since distances to safetywere not great. Exactly the samestrategy employed would likelyhave succeeded without incident afew hours earlier, or perhaps evenat that time of day if no suddenchange in wind velocity or direc-tion had occurred.

Had the spot fire not been in linewith one already controlled, or hadnot been hidden by the sharp littleridge in the bottom of ClaytonCreek, it would have had earlierattention from the Park crew fromTrail Ridge, and again the situationwould have been changed.

More time on the part of eitherClayton or Post to fully scout outthe potentialities of the fire aheadof the crews might have preventedthe tragedy.

Earlier arrival of the new crew,even by as little as a half hour,would have resulted in completingthe new line and would have con-centrated the attention of allsupervising officers and man-poweron all threats to holding it. Thiswould have resulted in a differentdistribution of the crews and prob-ably slight danger. Many otherpremises may also be drawn, butthe matter of timing of action ofthe fire vs. movement of men gave the distinctive and fatal combination.

Other Articles on the Blackwater FireGodwin, D.P. 1937. The handling of the Blackwater Fire. Fire Control

Notes. 1(6) [6 December 1937]: 373383.

Brauneis, K. 2002. 1937 Blackwater Fire investigation: Boost forsmokejumpers? Fire Management Today. 62(2): 2426.


Aerial view of Blackwater Fire, looking southeast directly up the head of Clayton Creek. Dashed line indicates thelost control line. Center arrow shows the gulch where Ranger Clayton and seven men were trapped. Arrowtoward the upper left shows the open point on the ridge where Ranger Post took refuge with his crew.


LESSONS FROM LARGER FIRES ONNATIONAL FORESTS, 1938*Roy Headley

Editors note: Roy Headley tookthe lessons learned from fieldreports on larger fires (fires cov-ering more than 300 acres [121ha]). In places, he added com-ments, indicated by brackets andthe italicized notation Ed.

Among fire control men there is areal difference of opinion on theextent to which study should beconcentrated on the larger fires asdistinguished from all fires. Somebelieve that concentration of atten-tion on large fires is dangerous andmay result in neglect of the meth-ods and principles which are ofimportance in keeping small firesfrom becoming larger.

The answer is probably a matter ofproper balance. Both classesshould be studied. Failure toreview the action taken on eventhe smallest fires invites weakness-es that will let more little onesgrow into big ones. But larger fireshave characteristics peculiar tothemselves. Management of thelarger fire jobs is a rather distinctbranch of controland a backwardbranch. Failure to study the actionon these larger fires invites weak-nesses that will let big ones growinto bigger ones. These bigger oneshurt, too. Out of a total of 219,000acres (88,000 ha) burned in 1938,35,000 acres (14,000 ha) were lost

* The article is an abridged version of sequential arti-cles published in Fire Control Notes 3(3) [Summer1939]: 617 and 3(4) [Fall 1939]: 3045.

When this article was originally published,Roy Headley headed the Division of FireControl, USDA Forest Service, Washington,DC.

in one fire. Seventy-five thousandacres (30,000 ha) were lost in thefour largest fires.

Lessons LearnedSuch an infinite variety of prob-lems are involved in the manage-ment of large fire jobs thatthoughtful men seldom fail tolearn from each one somethingwhich should be guarded againstin the future, something whichshould be done differently, somecherished belief which must bemodified or abandoned. For 35years I have been working on orobserving suppression jobs, but Istill learn something from everyfire I reach.

Sometimes, alas, we learn thesame lesson over and overor dowe? For example, I have learnedthroughout many years that thereis some flaw in our management oflarger fires which keeps us fromgetting a reasonable output of heldline from a crew of a given size.Plenty of other people have learnedthe same thing. But, untrained aswe are in the science and art ofmanagement, we have not foundways to act satisfactorily on whatwe have learned. Our learning hastoo often failed to lead to produc-tive action.

The first essential in such mattersis to grasp the need for change, thenature and importance of a prob-lem, the chance to introducesomething better. With that fact inmind, the outline for 1938 reportson larger fires requested a recordof lessons learned by the man ormen who had most to do with eachfire. Some of the most suggestiveanswers received are quoted in thisarticle. Perhaps these notes willhelp reduce the number of timeslessons have to be relearned bydifferent menor by the samemen.

Northern Region (R-1)AbsarokaChico fire429 acres(174 ha). The fire got over theline because of incomplete mop-upand men being gathered in abunch to get water. The foremanin charge was a young, inexperi-enced administrative guard who, inhis zeal to be helpful, left his crewto help on another part of the line.Had he had more experience, hewould have recognized that hiscrew lacked training and knowl-edge of fire fighting, and he wouldnot have left them. This might alsobe traced to the fact that the sectorwas too long for efficient handlingby the sector boss. The countrywas so steep that the man incharge could not cover his wholeline often enough.

Failure to study the action on larger fires invites weaknesses that will let big ones

grow into bigger ones.


KaniksuGoose Creek fire544acres (220 ha). An organizationprinciple was violated in thatauthority on the ground was divid-ed between two heads and coordi-nation was not effected. Two men,either qualified to handle theentire fire, were on the job. Eachassumed the other to be in charge.Each aggressively worked on oppo-site sides of the fire, expecting theother fellow to see that balancewas maintained.

The men planning the attack wereconfronted by a fire burning veryfast and throwing spot fires ahead.Fire was in a heavy, dry, white-pineslash area which had been loggedduring the summer and waspushed by a 20-mile (32-km/h)wind. It was figured that the firewould travel in a northeasterlydirection, which it did to a certainextent, but not as fast as was antic-ipated. Consequently, more menwere figured for the northeast sidethan actually needed. If our fore-sight had been as good as ourbacksight, more men would havebeen placed on the west side wherethere were barely enough men.

KootenaiRocky Creek fire380acres (154 ha). The old story thatan undetected spot fire blew upand caused this to be an extra-peri-od fire holds true here.

In an analysis of the action on thisfire by the supervisor, administra-tive assistant, two rangers, and analternate and dispatcher who wereon this fire, it was agreed that the25 men who were searching forspot fires were sufficient under the

conditions that existed at the timeof this fire. However, a spot fire gotaway, and caused the second run;therefore, the number of men orthe organization of them musthave been at fault.

Bulldozers were used to constructapproximately three-quarters of amile (1.2 km) of line on this fire.The main criticism was on accountof some of the burning fuel havingbeen covered with duff and dirtwhich added hours to the timerequired to do the mop-up job.

The time of arrival, the rate ofspread, and the size and value ofthe rapid-spread area determinedthe action to take. I believe a bull-dozer line under not too difficulttopographic conditions can be con-structed in less time than anyother machine we have on hand.The cost of mop-up is, of course, aproblem, but corral is the fist job.Because of the value of bulldozers,and the possibility of breakdowns,I recommend an extra machine ona fire-line construction job.

California Region (R-5)AngelesSan Antonio fire3,270 acres (1,323 ha). We wereslow in completing lines on somesectors, because of using crews toolarge20 men per experiencedcrew leader. We should have (forthis particular fire, and it is true inmost fall fires that spread witherratic lines and many spots) usedsmall crews, 5 men to each trainedfire fighter, and made sectors smallenough to provide closer supervi-sion and more effective perform-ance.

Splendid cooperation with theUnited States Weather Bureau intheir special fire-weather forecastsduring the entire fire season, andspecifically during the progress ofthis fire, made possible a moreaccurate planning program for firepersonnel needs and strategy to beemployed (particularly on backfirework).

Records show and this fire con-firms: Our late fall fires are a prob-lem we have not met in southernCalifornia. In the absence of earlyrains, in September and Octoberwe get Santa Anna conditions,extremely low humidity and highwind velocities up to 50 miles perhour (80 km/h). These conditionssometimes last a few dayssome-times 3 weeks. Practically all seri-ous fires in southern California forthe last 10 years have started underthese conditions, indicating wehave pointed our efforts to the nor-mal season conditions and gainedmuch ground, but now we mustpoint to these abnormal fall condi-tions and plan to meet them by:

1. Study more intensively, thebehavior of these late fall fires.

2. Provide a 24-hour patrol andlookout service. This was a nightfireas have been two or threeothers.

3. Make more intensive use of clo-sures regardless of private prop-erty interests.

4. Shorten elapsed time betweendiscovery and first action bymovement of equipment andmen so as to concentrate in highhazard areas and high occur-rence zones.

5. Intensify use of secondary look-out points.

6. Have night suppression crew onduty, dressed and ready to go.

For 35 years I have been working on or observing suppression jobs, but I still learn

something from every fire I reach.


In summary: Put additional heaton every control effort we normal-ly practice for regular summerseason.

KlamathRed Cap fire16,196acres (6,554 ha). The suppres-sion action on this fire was charac-terized by insufficient manpowerand overhead in the first five burn-ing periods, then a sudden build-up of men in a belated attempt toconform to the Foresters policy ofimmediate control. This suddenbuild-up led to placing more menin the field than could be ade-quately servicedwith consequentloss of efficiency. For example, ondivisions 1 and 2, zone B, it took400 men (including camp workers,etc.) 4 days to build and backfire404 chains (26,664 feet [8,127 m])of line. Theoretically, 400 menshould have built the line in 1 day,which would have been in accor-dance with the Foresters policy;but actually, they lost 50 percentefficiency because of lack of foodand bedding; and it is estimatedthat about half of the remainingeffort was dissipated in too wideline construction, so 100 men fullyserviced and adequately supervisedcould have accomplished the samejob in the same time.

The Weather Bureau fire-weatherfield station forecasts were usedthroughout the fire. The assuranceof this station that certain favor-able conditions would continuemade it possible to plan and con-struct a line into the head of theRed Cap Canyon, thereby savingabout 4,000 acres (1,600 ha) thathad appeared to be doomed.

The marvelous do or die, stick tothe bitter end spirit of the short-term force was magnificent. It waswith a feeling of deepest regretthat with the first big rain storm

on October 1 I was forced to dis-miss them with the expressed hopethat they could find enough to dothrough the winter to come backmentally and physically fit to tack-le another season.

San BernardinoArrowheadfire12,362 acres (5,003 ha).The Arrowhead fire originally start-ed in a cabin on the crest of theSan Bernardino Mountain range.After very thorough investigationas to the cause of the fire, it isbelieved that it was due to faultyflue construction. The defect wasbelieved to have been where thethimble fastened to the brickchimney.

Extreme fire weather conditionsexisted at the beginning of this fireand continued throughout. A highwind was blowing at gale force, 45to 50 miles per hour (7280 km/h),and the humidity was very low,about 7 percent.

The cabin in which the fire startedhad wooden, untreated shingles,also others that burned near thisone. Due to the high winds theburning shingles were probably thegreater cause of the many spotfires.

The rate of spread of this fire wasextremely great during the firstfew hours. Approximately 8,000acres (3,200 ha) burned during thefirst 7 hours or over 1,000 acresper hour (400 ha/h).

Because of the high rate of spreadit is doubtful that any means ofpreparedness in presuppression

would have reduced the size of thisparticular fire. However, it wasthought that we were deficient inthe number of tank trucks avail-able at the beginning of the fireand during the fire. A tank truck atArrowhead ranger station would bevery desirable and had there beenone there it would have arrived onthe fire about 7 minutes after thefire was reported instead of 36minutes as was the case. Eventhough there had been a tanktruck at Arrowhead ranger station,it is doubtful that it would havemade any difference in this fire.But as a presuppression lesson wecan see the importance of havingfavorable distribution of tanktrucks.

All line constructed and lost wasuncompleted line. All backfirework that was done was heldalthough it slopped over in places.Orchard torches were used. Noacreage was burned through back-firing which would not have beenlost by the fire anyway.

SequoiaFish Hatchery fire500acres (200 ha). Investigationshows that regardless of any sup-pression action taken after thecrews arrived on the fire, it couldnot have been controlled while stillsmall. This fire was simply movingtoo fast after it got up a headway. Itis barely possible that had we beenable to put about 50 men on thisfire at the time the first crewsarrived, the fire could have beenheld to a small acreage. However,there was no failure in first attackaction, as the first crew traveling10 miles (16 km), arrived 17 min-

The large area and rapid spread of this fire was due entirely to the dense and continuous

stand of cheat grass.


utes after the origin of the fire. Wefeel, therefore, that the only waythis fire could have been preventedfrom becoming large would havebeen for the district ranger to havetaken such action earlier in theseason as would have preventedthe fire from starting. This mighthave been possible had he specifi-cally designated a dumping groundfor the use of the Kern CountryJuvenile Camp and had thisground thoroughly fire-proofed atthe stated of the season.

The fire was in very rough countrywhere night work was difficult.District Ranger Stathem nowbelieves that it would have beenbetter business to have reduced hisnight crews in size, concentratingmost of his manpower on the lineduring daylight hours.

The lower fire line was built paral-leling an oiled road and in someplaces within 800 feet (240 m) ofthis road. The reason for this wasthat it was believed at the timethat it was necessary to do this toprotect a high-voltage transmis-sion line that was between the fireline and the road. Looking back onthis, it is easily seen, however, thatthere was an opportunity to dropback to the road and hold the fireand also protect the transmissionline with a much smaller crew,inasmuch as the line would havebeen easier to hold on the road.

ShastaSalt Creek fire1,690acres (684 ha). The most impor-tant lesson that I learned from thisfire was that it burned more rapid-ly at night, especially downhill,

than I had thought possible, andas a consequence, the fire made anadvance beyond the point where itwas estimated it would be held.

ShastaBig Lake fire745 acres(301 ha). The most importantthing I learned on this fire wasthat in slash type cover one shouldnever try to attack the head of afire if it reaches an area of morethan 3 acres (1.2 ha). To attack therear first and pinch it in on bothsides, putting the heaviest attackon the side toward which the windis blowing.

ShastaMount Hedron fire8,300 acres (3,400 ha). This fireoriginated in a grass sagejuniperfuel type in gently rolling countryat a time when the wind was blow-ing approximately 20 miles perhour (32 km/h). The fire wasattacked by 20 men and 2 tanktrucks within 11 minutes of itsorigin, and at that time it wasabout 3 acres (1.2 ha) in size andspreading rapidly. Cooperatorswere immediately called upon forhelp, but by the time they arrivedthe fire was completely out of con-trol and was heading north beforea strong tail win.

The action of the fire indicatedthat it probably would have been a300-acre (120-ha) fire even thoughthere had been a full camper ofCCC enrollees at Leaf which wasonly 8 miles (12 km) away. Itseemed to be one of those firesthat had to make its run.

TrinityGlennison Fire370acres (150 ha). This fire reached

the size it did as a direct result ofcrowning through the unburnedcanopy on the day following con-trol. Past records will show thatthis is a common occurrence onfires in the Canyon Creek areas,and a special effort had been madeto mop up the entire fire before theburning period of the day followingcontrol.

In spite of the efforts of 65 menwith backpack pumps and a powerpumper, the fire flared up frominside and crowned out in theunburned canopy of live oak.

If there is any lesson to be learnedfrom this experience, I believe it isthe necessity of having enoughmen with water equipment tocover the entire fire before thebeginning of the next burning peri-od. This is a rather difficult andcostly procedure, especially whencontrol is not achieved before day-light. The flare up on this firebegan at 12 noon, which did notallow much time to cover theentire fire. Also, it is not advisableto send men into the burn before ithas had time to cool sufficiently toallow them to work safely.

Control of this fire the day follow-ing was achieved by following theabove procedure to a large extent.However, weather conditions weremore favorable until the late after-noon, which allowed more time tomop up.

In summarizing the above, perhapsthe most important point is toachieve control as early as possiblethe first night so that the fire willcool down before daylight, andallow more time for mop up.

TrinityLittle Bear Wallow Fir2,200 acres (890 ha). CivilianConservation Corps (CCC)

Records show and this fire confirms: Our Santa Ana fires are a problem

we have not met in southern California.


enrollees are ineffective on a firewhen not thoroughly trained in firesuppression, and they must beaccustomed to hard hiking inrugged country and able to doeffective work after arriving at thefire. In the case of this fire, the ini-tial attack was made by an experi-enced guard and two greenenrollees. The boys, who weretransferred here with a New Yorkcompany a day or so previously,were soft, untrained, and notaccustomed to hiking, and not onlyslowed down the guards traveltime to the fire, but were uselessafter arriving there.

North Pacific Region(R-6)SiskiyouCedar Camp (Chetco)Fire34,627 acres (14,014 ha).Something must be done to stopincendiary fires from being startedto keep large fires burning. On thedivision in which I worked, it is apositive fact that one fire was set,and I am relatively sure that otherswere set in the same area. Take thelocal mans advice as to fire behav-ior and wind drafts if he is reliable.Reliable local men are indispensa-ble as scouts, especially in unfamil-iar country.

SiskiyouNome Fire5,800acres (2,300 ha). Planes proved tobe an excellent means of supplyingfire camps both from the stand-point of speed and economy. Campswere serviced that otherwise wouldhave had to be supplied by manpack, as was done for several dayswhen planes were grounded, whichrequired 8 hours of hard packingby good men to get 40 pounds ofsupplies into camp.

Planes will not eliminate the use ofhorses. During the time that thisfire was burning the smoke becameso dense the planes were grounded

for several days, which clearlybrought out the fact that packstock must be relied upon at times.

More use should be made of theindirect method of control on firesof this size and in this type oftopography. There is always thehuman aversion to burned areathat is continually cropping up onfires of this kind.* Men work tooclose to the fire, thus sacrificingtime and labor. More satisfactoryresults can be obtained by drop-ping back and taking advantage ofnatural breaks in topography.

A stream, unless it is large andclear of brush, makes a poor fireline. The variable winds up anddown streams, the possibility ofrolling logs and falling snags, andthe rapid spread of a fire up thesloe once it crosses a stream,makes it exceedingly dangerous touse.

FremontBonanza Fire9,155acres (3,705 ha). The large areaand rapid spread of this fire wasdue entirely to the dense and con-tinuous stand of cheat grass(Bromus tectorum) with very littletimber overstory. Consequently,the usual forest fire standards arenot fully applicable. However, thefollowing was learned about thisparticular country.

1. Must keep roads cleared as firebreaks.

2. Must construct additional crossroads.

3. Must make definite plans forsuppression (outline prepared).

4. Overhead, in general, needsmore mop up training.

5. CCC enrollees must be shownhow by fire foremen and strawbosses.

6. On very large fires it is better touse machinery for trench con-struction at the sacrifice ofsome acreage than to depend onhand work too close to the fires.

Investigation of the cause of break-over on August 26 showed thatwhile a good trench and cleanburn had been obtained on theextreme southeast corner of thefire, and mop up had been carriedback of the line an adequate dis-tance for ordinary circumstances,the crowning of a small green pinetree on a low ridge about 300 feet(90 m) inside the line threw sparksinto the thick cheat grass andjuniper timber which quicklycrowned and spread under thebrisk breeze. Men working in thevicinity of the break took actionalmost immediately, but wereunable to cope with the heat of thecrowning junipers and did nothave time to use the indirectmethod.

Rocky MountainRegion (R-2)RooseveltJenny Lind Fire664acres (269 ha). I believe the most

* Editors note: American Indians routinely used firefor natural resource management, a practice that someEuropean settlers adapted to their needs. But by the1930s, the culture of fire control was so deeplyingrained that aversion to burned area seemed quin-tessentially human.

An important principle is that the front of a fast running fire is often untouchable. But a lull always comes.

The job is to be sure to grab it for keeps during that first lull.


important point to be noted is thegreat difficulty of insuring ade-quate initial action on a veryaggressive fire in its initial stages.[Three miners, 100 feet (30 m)away, saw fire and attacked it 5minutes after it started. Districtranger with one man arrived at fire35 minutes after it started. Butlarger crews did not reach fire intime to prevent 39- to 54-mile(6387 km) wind from starting fireon its first run. Probably mostimportant lesson to be learned isthat on a forest not provided withregular detection and other facili-ties common to western fireforests, a fire starting in slash in agale of wind could be and wasstopped at 664 acres (269 ha). Ed.]

Almost uniformly there was toomuch tendency to bury burningmaterial along the fireline in sucha way that it only prolonged theburning and the mop-up period.The correct use of dirt has been somuch emphasized that there arefew if any of the foremen ignorantof the undesirability of partiallyburying heavy material; but thereseemed to be great difficulty ingetting thus particular fault cor-rected.

A common fault which was per-haps the most serious was that ofinclosing ragged fire perimeterinside a line without taking anymeasures to clean burn. Thisneeds to be given a great deal ofemphasis on all forests.

Southwestern Region(R-3)GilaLookout Ridge Fire575acres (233 ha). Class 6 burningconditions prevailed and were notrecognized. The fuels were verydry and the wind was strong. Thelesson learned here is that a firedanger station would have fur-nished a definite check on burningconditions, and this could thenhave been promptly followed byproper strengthening of the guardforce to meet class 6 or emergencyconditions. [This forest is now sup-plied with a fully equipped station.Danger stations and danger ratingsare designed to avoid just this sortof failure to recognize changes infire danger. Ed.]

Poor line location due primarily totrying to hold fire to a minimumacreage with subsequent loss ofexcess acreage was an importantfactor. This is merely a matter oftraining in suppression. [Yes, buttraining in judgment, which is themost difficult and most backwardfield of training. Ed.]

GilaIron Creek Fire2,318acres (938 ha). Here again class 6burning conditions existed unrec-ognized. Too much line was lost byworking too close to the fire.[Understand that this fire and thisweakness were used by the regionfor training by the case method.Ed.] Width of fire lines must begoverned by (a) fuel on ground; (b)steepness of slope; (c) wind veloci-ty. In the past, fire lines have been

entirely too narrow for safety inhandling backfiring in windyweather. [Importance of trainingin judgment again emphasized.More common error is to makefire lines wider than need be forbackfiring. Ed.]

Intermountain Region(R-4)WasatchShepherd Creek Fire960 acres (388 ha). Particularlysignificant is that the fire, carriedby a high wind, spread over 700acres (280 ha) within 2 hours afterit started. It traveled so fast thatthe fire truck, plow unit, and mencould not keep pace. Forty menwere on the fire within 20 minutesafter it started, and, with our firetruck which suppressed a mile (3.2km) of fire edge, and the plow unitthat made over 3,000 feet (900 m)of fire line, it was impossible tocope with the high rate of firespread. [But the wind dropped andthe fire died down, after which thecrews had a chance and did notmuff it. An important principle isthat the front of a fast running fireis often untouchable. But a lullalways comes. The job is to be sureto grab it for keeps during thatfirst lull. Ed.]

Southern Region (R-8)ConecuhBig Fire576 acres(233 ha). In the attack on thisfire, as in most other fires on theConecuh, (wire-grass type ofground cover), everything depend-ed upon the success or failure ofthe tank truck. From the time ofattack with this truck until thewater supply was exhausted (15minutes) 19 chains (1,056 feet[321 m]) of line were built. A reor-ganization has been made to pro-vide for replenishment of the sup-ply of water for the tank truck.

To summarize, it is apparent that extreme effort should be made

to satisfy the requirements of the Foresters policy as to control in

the early burning stages.


A whirlwind picked up burningmaterial from as far as 100 yards(90 m) inside the line and droppedit outside in the rough, resultingimmediately in a break-over oflarge proportions. On at least twooccasions whirlwinds hit the fireline while suppression was beingcarried on with tank trucks,resulting in loss of sort lengths ofline and excessive use of water.Variability of wind direction andprevalence of whirlwinds are fac-tors not ordinarily considered indanger rating.

Evidence indicates there was timeto have cut a line and backfiredacross the head. This would havespeeded up control and promisedbetter chance of holding to asmaller acreage.

ConecuhBradley Fire792acres (321 ha). There is a definiteneed for a mobile tractor-plowsuppression unit, capable of rapid-ly plowing a line from which back-firing, could be done. This type ofequipment appears to offer theonly practicable means for stop-ping the head of the occasionalbad fire short of natural breaks.

The relative humidity on the dayof the fire (December 3) did notreach a point that could be calledlow. This shows that relativehumidity, in itself, cannot berelied upon as an indicator of firedanger.

OcalaPleasant Flat Fire2,161acres (875 ha). The ground ofthe timber type through whichthis fire burned is generally cov-ered with water during periods ofnormal weather, but dries out sev-eral times a year so that itbecomes inflammable. Since it isnot very accessible by tank trucks,

the construction of permanentmeandering 8-foot (2.4-m) fire-breaks on the higher ground with-in the type will prove very helpfulin suppression of any future fireswhich may occur.

Summarizing, I would say that themost important thing this fireshowed was the inability of theranger to make efficient use of themanpower and equipment that wasavailable at the time of the fire.The matters of new and betterequipment are secondary. [Theranger may take comfort from thefact that many others, includingbig shots in fire control, sufferfrom the same inabilitybut thisshould not deter him from seekingto be mentally prepared for effi-cient management of the next fire.Ed.]

KisatchieSlash fire407 acres(165 ha). Two hundred and tenacres (85 ha) of the fire had onlyrecently been planted to slash pine.The slash pine of course was totallydestroyed. The rough was com-posed of wire grass, broomsedge,and a fairly heavy carpet of hard-wood leaves. The rough was of 5years accumulation, and can beconsidered heavy for longleaf type.

Investigation revealed that the firewas handled very well by the look-out men and the dispatcher. Thediscovery, report, and getawaytimes are excellent. However, thedecision of the superintendent tosend only 6 men with the juniorforeman in a pickup to check on

the fire was sadly at fault. The daywas not an extremely bad one;however, a fresh south wind wasblowing and any fire located withina quarter or a half mile (0.40.8km) south of the boundary certain-ly should have merited both thefire truck (tank truck and Panamapump) and a full 16-man crewwith the standard fire-fighting toolbox. If the junior foreman had hada full 16-man crew and fire truckhe would have unquestionably heldthe fire with only an acre (0.4 ha)or so of national forest landburned.

DeSotoSaucier Fire635 acres(257 ha). The weather conditionsmade the hazard extreme. Therehad been no rain for 3 days. Therelative humidity was 33 at noon.Wind velocity varied from 8 to 16miles per hour (1326 km/h) dur-ing the fire. The ground cover wasbroomsedge grass. The flames hada range of 40 to 60 feet (1218 m),crowning at times and spottingahead.

Instructions were to backfire froma fire lane on the west side. Thisfirebreak had been constructed 2years ago and had not been main-tained since this time. As a result,it was grown over and offered noeffective barrier from which tobackfire. The extreme weatherconditions made it improbable thateven if the firebreak had beenrecently maintained, the crewwould have been able to hold theirbackfire on this narrow line.

The most important lesson that I learned from this fire was that it burned more rapidly

at night, especially downhill, than I had thought possible.


DeSotoHell Hole Creek Fire509 acres (206 ha). This fireemphasized how immediate andcontinued work on the flanks of alarge fire can keep the acreage to aminimum and prevent new headsfrom forming when the wind shifts.Crews worked the flanks continual-ly so that when the head was final-ly stopped the entire fire was cor-ralled within a very few minutes.

The head of this fire presented awall of flames 20 feet high (6 m),with a range of 30 to 40 feet (912m). It was burning in a 5-yearrough. Several of the fire fightershad narrow escapes while attempt-ing to stop the head. Two differentpump men were forced to abandontheir pumps in order to outrun thefire.

This fire, as well as other largeones on the DeSoto, shows that inextreme weather conditions a greatdeal depends on the dispatchersjudgment in placing crews, notonly for the head but for work onthe flanks. The speed at whichthese fires travel offers little oppor-tunity to correct tactical mistakeson the fire line.

North Central Region(R-7)GardnerTower Fire489 acres(198 ha). This was one of a seriesof 25 fires burning on the districton this date. The fire danger wasclass 6 and the wind velocity was20 miles per hour (32 km/h). Thefire spread rapidly in highlyinflammable leaf and grass fuel.

Topography was rugged with steepslopes, which accelerated the rateof spread. Winds shifted frequentlyand small whirlwinds occurredalong the fire front. As a result thefire had a number of heads, withone traveling along the top of eachmain ridge. The fire advanced in asolid sheet of flame, leaping 6 feet(2 m) high in most places, and themen could not get near the firebecause of the intense heat. Spotfires, jumping in many cases over100 feet (30 m) ahead of the mainfire, were common. There was 2 to3 years accumulation of leaves onthe area.

When Assistant Ranger Barryarrived, he took charge of the fireimmediately. He made a quicksize-up of the situation, deter-mined that direct attack was inef-fective [It had already failed. Ed.],

A whirlwind picked up burning material from as far as 100 yards inside the line and

dropped it outside in the rough, resulting immediately in a break-over of large proportions.

and directed the crews to dropback several hundred feet from thefire edge to construct line. Thetechnique used, which consisted ofraking a line and backfiring fromit, was considered satisfactoryunder the circumstances. Thisdropping back was essentialbecause whirlwind conditionscaused fires to spot 100 to 200 feet(3060) ahead of the main fire.

Knowing the conditions as areknown now, fire managementcould have been improved by back-firing from the truck trail andkeeping the fire from crossing it.However, the use of this techniquewould have required at least 25additional men as there were sepa-rate heads traveling toward theroad.

The general belief that all fires inMissouri hardwoods can be han-dled by direct or parallel methodsis false. On certain days, conditionsare such that the only practicalmethod is to drop back to naturalor cultural barriers, such as theroad in this case, and sacrificesome burned acreage in order toinsure control.

A more extensive use of watershould be resorted to on such days.On this fire three men could notcontrol a 5-foot (1.5-m) spot fireby use of rakes. A marine pumpwould have saved the first linebuilt on this fire.


LESSONS FROM LARGER FIRES ONNATIONAL FORESTS, 1939*Roy Headley

Editors note: Roy Headley tookthe lessons learned from fieldreports on larger fires (fires cov-ering more than 300 acres [121ha]) in the Eastern, Southern, andNorth-Central Regions. He indicat-ed the source following eachreport. His own summaries andcomments are indicated by brack-ets and the italicized notationEd.

Eastern Region (R-9)George WashingtonChestnutRidge Fire872 acres (353 ha).This fire on March 9, under veryserious burning conditions, wasprobably caused by children livingin the vicinity. The fire caught thepreparedness organization of thedistrict napping. Towers were notmanned. Weather stations and dis-trict fire truck not in operation,and the district ranger not intouch with his dispatcher. Failureto recognize emergency conditionsand to build organization on thebasis of emergency conditionsresults in fires of this class.

Once organized the suppressionaction was satisfactory. However,because of lack of preparedness itrequired 3-1/2 hours to get thesuppression organization to thefire. The result was a large fire, andthe lessons learned by all con-cerned are that regardless of previ-ous weather, conditions may

* The article is an abridged version of an article pub-lished in Fire Control Notes 4(1) [Winter 1940]: 1517.

When this article was originally published,Roy Headley headed the Division of FireControl, USDA Forest Service, Washington,DC.

The Chestnut Ridge fire showed that weatherconditions may rapidly change during early spring,

and that the field organization must be kept on its toes.

change very rapidly during theearly spring, and that the fieldorganization must be kept on itstoes. Unsigned forest report

A complete failure of districtranger to recognize degree of firedanger and plan accordingly.Corrective action has been taken.The most valuable lesson in thiscase involves the continuous dailyuse of fire-weather stations todetermine degree of danger follow-ing rain. Very rapid changes involv-ing only a few hours from no dan-ger to high danger are to beexpected during March, April, andMay. R.M. Evans, regional forester

Southern Region (R-8)De Soto, Leaf River DistrictPlantation No. 24 Fire1,682acres (681 ha). [When discovered5 minutes after known time of ori-gin, this incendiary grass fire was 3 acres (1.2 ha). Report and get-away took 5 minutes. Three miles(4.8 km) of auto travel, 10 minutes.When reached by the first crew of17 men, 20 minutes after origin,the fire was 20 acres (9 ha). But itburned 1,682 acres (681 ha),including 870 acres (352 ha) ofplantation, before it was corralled10 hours and 25 minutes after origin. In one place, wind velocity

is reported as 2 to 19 miles (3.231 km/h); in another, 25 to 38 miles (40-61 km/h). Five dayssince last rain of 0.28 inch (0.71cm). Number of chains per manhour, 2.3 (152 feet [46.3 m]). Muchconfusion and disorganization onthe line. Maximum number of menengaged, 65. Ed.] The initial tac-tics used by the fire boss were cor-rect for this fire, and had the heavyequipment functioned properly,the fire would have been held toless than 100 acres (40 ha).

De Soto, Leaf River DistrictLeafNo. 35 Fire416 acres (168 ha).[Fire was 75 acres (30 ha) whendiscovered 2 hours and 17 minutesafter guessed time of origin. Itspread 25 acres (10 ha) during the53 minutes required for report,get-away, and the travel. Ed.]

Low visibility accounts for the sizeof fire when discovered. Therewere five sets and the purpose ofthe fire was to provide sheeprange. The exposed ridge site andhigh winds made suppression diffi-cult. A fire usually occurs in thislocation about the same time eachyear. Next year it is planned topatrol these areas closely in aneffort to prevent the fire or catchthe incendiary. V.B.McNaughton, fire assistant


De Soto, Biloxi Ranger DistrictUniversity No. 28 Fire1,238acres (501 ha). [Another grass,stockman, incendiary fire but withonly one set. This time the lookoutman snapped in with discoverywithin 4 minutes after guessedtime of origin, when the probablearea was one-tenth acre (0.04 ha).But report took 42 minutes. In the65 minutes required for report,get-away, and 9 miles (14 km) autotravel, the fire spread to 350 acres(140 ha). Wind velocity reported as13 to 18 miles per hour (2129km/h) when fire was first reachedand during biggest run. Ed.]

De Soto, Biloxi Ranger DistrictCamp Branch No. 16 Fire419acres (170 ha).

wildland fire behavior case studies and analyses: part 1 wildland fire ...

Documents

Transcript of wildland fire behavior case studies and analyses: part 1 wildland fire ...