2008 CFFDRS Weather Guide.pdf

8/10/2019 2008 CFFDRS Weather Guide.pdf

1/87

B.D. Lawson and O.B. Armitage

Weather Guide

Canadian Forest Fire

Danger Rating System

for the


2/87

The Northern Forestry Centre is one of five centres of the Canadian Forest Service, which has its headquarterin Ottawa, Ontario. This centre undertakes the regional delivery of national projects.

The Canadian Forest Services main objective is research in support of improved forest management foeconomic, social, and environmental benefits to all Canadians.

Le Centre de foresterie du Nord constitue lun des cinq tablissements du Service canadien des fortsdont ladministration centrale est Ottawa (Ontario). Le Centre entreprend la ralisation rgionale de projetnationaux.

Le Service canadien des forts sintresse surtout la recherche en vue damliorer lamnagement forestieafin que tous les Canadiens puissent en profiter aux points de vue conomique, social et environnemental.


3/87

WEATHER GUIDE FOR THE

CANADIAN FOREST FIRE DANGER RATING SYSTEM

B.D. Lawson1and O.B. Armitage2

Canadian Forest Service

Northern Forestry Centre

2008

1, 2Ember Research Services Ltd., 4345 Northridge Cres., Victoria, British Columbia V8Z 4Z4.


4/87

Her Majesty the Queen in Right of Canada, 2008

Natural Resources CanadaCanadian Forest ServiceNorthern Forestry Centre5320 - 122 StreetEdmonton, Alberta T6H 3S5

Catalogue No. Fo134-8/2008E-PDFISBN 978-1-100-11565-8ISSN 0831-8247

For an electronic version of this report, visit the Canadian Forest ServiceBookstore at http://bookstore.cfs.nrcan.gc.ca/

TTY: 613-996-4397 (Teletype for the hearing-impaired)ATS: 613-996-4397 (appareil de tlcommunication pour sourds)

Library and Archives Canada Cataloguing in Publication

Lawson, B. D.Weather guide for the Canadian Forest Fire Danger Rating System[electronic resource] / B.D. Lawson and O.B. Armitage.

Includes bibliographical references.Electronic monograph in PDF format.

ISBN 978-1-100-11565-8Cat. no.: Fo134-8/2008E-PDF

1. Fire weather--Canada.2. Forest meteorology--Canada.3. Forest fire forecasting--Canada.4. Fire risk assessment--Canada.5. Forest fires--Canada--Prevention and control.I. Armitage, O. B.

II. Northern Forestry Centre (Canada)III. Title.

SD421.37 L38 2008 363.370971 C2009-980001-2


5/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM III

Lawson, B.D.; Armitage, O.B. 2008. Weather guide for the CanadianForest Fire Danger Rating System. Nat. Resour. Can., Can. For.Serv., North. For. Cent., Edmonton, AB.

ABSTRACT

The Weather Guide for the Canadian Forest Fire Danger Rating System is intended primarily for operational wildland fire management personneland forest fire weather practitioners responsible for gathering, processing,and forecasting fire weather information in support of safe and effectivesuppression and use of fire. Accurate and representative weather observationsthat meet prescribed standards and specifications are necessary for accurateand representative calculation of all components of the Canadian Forest FireDanger Rating System. Weather-dependent components or modules arecalculated or computed for effective use of the systems two main subsystems,the Canadian Forest Fire Weather Index (FWI) System and the CanadianForest Fire Behavior Prediction (FBP) System. This weather guide includesdetailed specifications for locating and instrumenting fire weather stations,taking weather observations, and overwintering the Drought Code componentof the FWI System. The sensitivity of the FWI System components to weatherelements is represented quantitatively. The importance of weather that is notdirectly observable is discussed in the context of fuel moisture and fire behavior.Current developments in the observation and measurement of fire weatherand the forecasting of fire danger are discussed, along with the implicationsfor the reporting of fire weather of increasingly automated fire managementinformation systems.

RSUM

Le Guide sur les conditions mtorologiques de la Mthode canadiennedvaluation des dangers dincendie de fort (MCEDIF) sadresse principalementau personnel charg des oprations de gestion du feu en fort et aux spcialistesde lindice fort-mto chargs de la collecte et du traitement de linformationsur les conditions mto propices aux incendies de fort et de ltablissementde prvisions lappui dactivits scuritaires et efficaces de suppression etdutilisation du feu. Il faut disposer dobservations mtorologiques prciseset reprsentatives, conformes aux normes prescrites et aux spcifications,pour effectuer des calculs prcis et reprsentatifs de toutes de toutes lescomposantes de la Mthode canadienne dvaluation des dangers dincendie defort. Les lments ou modules tributaires des conditions mto sont calculsde manire permettre lutilisation efficace des deux composantes principales

de la MCEDIF, soit la Mthode canadienne de lindice Fort-mto (IFM) et laMthode canadienne de prvision du comportement des incendies (PCI) de fort.Le prsent guide expose notamment en dtail la marche suivre pour localiseret instrumenter des stations mtorologiques, effectuer des observationsmto et ajuster lindice de scheresse de la Mthode IFM en fonction desprcipitations hivernales. La sensibilit des composantes de la Mthode IFM auxlments mtorologiques est reprsente quantitativement. Le guide traitede linfluence des conditions mto non directement observables sur lhumidit


6/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM IV

du combustible et le comportement du feu. Il fait tat des progrs en matiredobservation et de mesure des conditions mto propices aux incendies et deprvision du danger de feu ainsi que des incidences sur la communication desconditions mto propices aux incendies de lautomatisation grandissante dessystmes dinformation sur la gestion du feu.


7/87


8/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM VI

WEATHER NOT DIRECTLY OBSERVABLE . . . . . . . . . . . . . . . . . 36

Vertical Structure of the Atmosphere . . . . . . . . . . . . . . . . 36Low-Level Jet . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Crossover . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38Dew and Frost . . . . . . . . . . . . . . . . . . . . . . . . . . 40

SENSITIVITY OF CFFDRS COMPONENTS TO WEATHER CHANGES . . . . . 41

Sensitivity of Fuel Moisture Codes to Weather Elements . . . . . . . 41Fine Fuel Moisture Code. . . . . . . . . . . . . . . . . . . . . 41Duff Moisture Code and Drought Code . . . . . . . . . . . . . . 43

Sensitivity of Initial Spread Index and Predicted Rate ofSpread to Wind Speed . . . . . . . . . . . . . . . . . . . . . 45

Sensitivity of Fire Weather Index and Predicted Fire Intensityto Wind Speed . . . . . . . . . . . . . . . . . . . . . . . . . 47

FIRE WEATHER FORECASTING . . . . . . . . . . . . . . . . . . . . . 48

Fire Weather Forecasts . . . . . . . . . . . . . . . . . . . . . . 49Fire Danger Forecasts . . . . . . . . . . . . . . . . . . . . . . . 49Fire Behavior Forecasts . . . . . . . . . . . . . . . . . . . . . . 50Implications of Fire Weather: Fire Management Information

Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51LITERATURE CITED . . . . . . . . . . . . . . . . . . . . . . . . . . 53

APPENDIXES

The Beaufort scale for estimating 10-m open wind speeds1. . . . . . 57

Comparison of hourly and diurnal models for the fine fuel2.moisture code . . . . . . . . . . . . . . . . . . . . . . . . 59

Latitude considerations in adapting the Canadian Forest Fire3. Weather Index System for use in other countries . . . . . . . . 67

FIGURES

Structure of the Canadian Forest Fire Danger Rating System (CFFDRS).1. 1

Structure of the Canadian Forest Fire Weather Index System.2. . . . . 2

Schematic diagram for calculating the six standard components3.of the Canadian Forest Fire Weather Index System. . . . . . . . . . 6

Daily patterns of temperature, relative humidity, and wind speed4.in July for a typical continental station, Leighton Lake, B.C.. . . . . . 7

Reduction of surface wind speeds according to roughness of5.surrounding terrain. . . . . . . . . . . . . . . . . . . . . . . . 17

Fire weather station in large clearing on open level ground..6. . . . . 18

Stevenson screen, on an open framework stand that is staked7.to the ground. . . . . . . . . . . . . . . . . . . . . . . . . . . 19


9/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM VII

Portable electric fan psychrometer, with battery-powered fan8.and wet-and dry-bulb thermometers. . . . . . . . . . . . . . . . 20

Hygrothermograph mounted in separate screen with9.double-louvered sides and double roof. . . . . . . . . . . . . . . 21

Standard anemometer height: 10 m, if the clearing is large10.enough that nearest timber edge is a distance of at least 5 timesthe height of the trees away from the mast. . . . . . . . . . . . . 21

Anemometer height adjusted for uneven ground or brush.11. . . . . . 22

Automatic fire weather station showing typical configuration of12.sensors, power supply, data storage, and communication. . . . . . 24

Placement of fire weather stations in hilly or mountainous terrain.13. . 28

Quick-deploy automatic weather station with nonstandard-height14.anemometer requiring wind speed adjustments. . . . . . . . . . . 31

Calibration curves for forest floor moisture content as a function of15.Drought Code (DC): national standard, coastal British Columbiacedarhemlock (CWH) forests, southern interior British Columbiaforests, and southern Yukon white spruce forests. . . . . . . . . . 35

Diurnal trends of temperature and relative humidity in which for16.a portion of the day the relative humidity is equal to or less thanthe air temperature, termed crossover. . . . . . . . . . . . . . 39

Sling psychrometer from belt-mounted weather kit used for local17.measurements of wet-bulb and dry-bulb temperatures andcalculation of relative humidity. . . . . . . . . . . . . . . . . . . 40

Effects of todays noon temperature on Fine Fuel Moisture18.Code (FFMC) for three levels of yesterdays FFMC. . . . . . . . . . 42

Effects of todays noon relative humidity on Fine Fuel Moisture19.Code (FFMC) for three levels of yesterdays FFMC. . . . . . . . . . 42

Effects of todays noon wind speed on Fine Fuel Moisture Code20.

(FFMC) for three levels of yesterdays FFMC.. . . . . . . . . . . . 43

Change in yesterdays Duff Moisture Code (DMC), given todays21.noon temperature. . . . . . . . . . . . . . . . . . . . . . . . . 44

Change in yesterdays Duff Moisture Code (DMC), given todays22.noon relative humidity. . . . . . . . . . . . . . . . . . . . . . . 44

Change in yesterdays Drought Code (DC), given todays noon23.temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Fire Weather Index (FWI) Initial Spread Index (ISI), Fire Behavior24.Prediction (FBP) ISI, and rate of spread for fuel type C-3 (maturejack or lodgepole pine) as a function of wind speed. . . . . . . . . 46

Fire Weather Index (FWI) and fire intensity for fuel type C-325.

(mature jack or lodgepole pine) as a function of wind speed. . . . . 48


10/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM VIII

TABLES

Physical properties of forest floor layers associated with the fuel1.moisture codes of the Canadian Forest Fire Weather Index System . . 3

Recovery of Fine Fuel Moisture Code (FFMC) after rain with three2.levels of temperature (temp.), relative humidity (RH), or windspeed (WS), with starting FFMC of 70 . . . . . . . . . . . . . . . . 6

Effect of wind on Initial Spread Index (ISI)3. . . . . . . . . . . . . . 8

Threshold rain values for fuel moisture codes4. . . . . . . . . . . . . 8

Effects of rain on the fuel moisture codes5. . . . . . . . . . . . . . . 9

Average hourly rates of change of temperature, relative humidity6.(RH), and wind speed at noon local standard time for select stations 10

Recommended height of anemometer for small clearings7. . . . . . . 22

Wind speed adjustment factors for anemometer mast height less8.than 12 m. . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

User-selected values and criteria for equation 2 constants a and9.b, overwintering the Drought Code . . . . . . . . . . . . . . . . 33


11/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM IX

FOREWORD

For nearly 30 years the publication Weather inthe Canadian Forest Fire Danger Rating System A User Guide to National Standards andPractices(Turner and Lawson 1978) has servedas the primary reference for the collection of fireweather data in Canada.

Although the basic weather inputs into theCanadian Forest Fire Danger Rating System(CFFDRS) have changed little, data collectiontechnology has moved forward significantly overthe years. In addition, continued research andtechnical observations related to the CFFDRShave increased our understanding of theinteractions among weather, forest fuels, and firebehavior, which form the basis of the system.

These observations have been published in avariety of documents. As a result, it has becomeincreasingly difficult for practitioners to keepabreast of recent developments and to readilyaccess the information that could be useful tothem.

In 2006 the Forest and Fire MeteorologyWorking Group, operating under the mandateof the Canadian Interagency Forest Fire Centre,proposed to update the 1978 document to bringit into line with current technology and practicesand to reflect recent scientific findings. Inaddition, the group sought to expand the scopeof the guide so that it would become a generalreference for issues relating to fire weather andfuel moisture as they apply to the CFFDRS.

The goal was to produce a consolidated,up-to-date reference for this material, whichwould be useful for both new and experiencedpractitioners, as well as other interested parties.Consequently, the document has been published

electronically in a form that will facilitate theincorporation of updates and new findings asthey become available.

Forest and Fire Meteorology Working Group,Canadian Interagency Forest Fire Centre, 2007

PREFACE

The Canadian Forest Fire Danger RatingSystem (CFFDRS) encompasses a numberof publications that document equations andinterpretive material defining and describing thisnational danger rating system, which have beenapproved by the Canadian Forest Service (CFS).At present, these publications are physicallyavailable in a three-ring binder titled CanadianForest Fire Danger Rating SystemUsersGuide.1

The publication covering fire weather mattersrelated to the CFFDRS that is currently includedin the CFFDRS users guide is titled Weather inthe Canadian Forest Fire Danger Rating System A User Guide to National Standards andPractices, by J.A. Turner (deceased 1979) and

B.D. Lawson (former member of the CFS FireDanger Group, retired from CFS in 1996). Thisregional publication (CFS Information ReportBC-X-177, 1978; also available in French) hasserved as a national weather guide for manyyears, but has gradually become outdatedbecause of technological and scientific changesin the collection and management of weatherdata, the computation of components of thedanger rating system, and some of the CFFDRScomponents themselves. An abbreviatedversion of the 1978 publication was included as

Chapter 12 (in Part B, Forest Fire Meteorology) ofEnvironment Canadas Forest Fire Management:Meteorology A Training Manual (EnvironmentCanada 1987).

1Canadian Forestry Service. 1987. Canadian Forest Fire Danger Rating System Users Guide. Can. For. Serv., Fire Danger Group. [Ottawa,

ON.] Three-ring binder (unnumbered publ.).


12/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM X

Considerable work toward a national weatherguide was done by the CFS in 1984, withextensive suggestions for revisions to BC-X-177being submitted by M.E. Alexander, NorthernForestry Centre, and others. In the preparationof the new weather guide, B.D. Lawson obtainedthese paper files from CFS and reviewed them,along with new comments from B.M. Wotton,Great Lakes Forestry Centre.

Like its predecessor, the new weather guidewill assist those responsible for establishing fireweather stations for danger rating networksand operating short-term applications such ascampaign wildfires2 and prescribed burns. Thisweather guide specifies the standards requiredfor the basic fire weather observations that areused to calculate components of the CanadianForest Fire Weather Index (FWI) System and theCanadian Forest Fire Behavior Prediction (FBP)System, the two principal subsystems of the

CFFDRS.

The weather guide also provides proceduresand adjustments for nonstandard weathersituations and weather-recording practices thatshould be of interest to all fire weather observers.The weather guide does not, however, includean availability list for weather instrumentationavailability, nor is it a detailed instruction manualfor taking weather observations with specificinstruments.

Although this weather guide references theGuide to Agricultural Meteorological Practicesof the World Meteorological Organization (WMO1968), it deviates in one important respect fromthe WMO-recommended standard for a clearingsize for a forestry weather station, as follows.

In its original publication, the WMO (1968)stated that, current information suggests thatthe diameter of the clearing should be at least 10times, preferably 20 times the tree height of thesurrounding forest. In a 1982 supplement (WMO1982), the WMO revised its recommendation onclearing size, stating that to minimize the effectof forest vegetation on air flow, the anemometermast should be located in the centre of anopening in the forest with diameter of at least 20times the height of the surrounding trees. Thislarger recommended clearing size was retainedby WMO in its 1993 forestry supplement (WMO1993). However, for continuity with established

CFS practice, the current weather guide retainsthe long-standing recommendation for forestfire weather station clearings in Canada that theanemometer be located in the centre of a clearingwith diameter at least 10 times the height of thesurrounding trees.

2A campaign fire is a fire of such size, complexity and/or priority that its extinction requires a large organization, high resource

commitment, significant expenditure, and prolonged suppression activity (synonym: project fire). Merrill, D.F.; Alexander, M.E. 1987.

Glossary of forest fire management terms. 4th Ed. Natl. Res. Counc. Can., Can. Comm. For. Fire Manag., Ottawa, ON. Publ. NRCC 26516.


13/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM XI

ACKNOWLEDGMENTS

The authors acknowledge Nick Nimchuk,Alberta Sustainable Resource Development,Forest Protection Division, and Eric Meyer,

British Columbia Ministry of Forests and Range,Protection Program, who spearheaded the effortto update the old weather guide. Together, theyinitiated support for the project at the nationallevel and secured the necessary contract fundingfrom their respective provincial fire managementagencies. A project to update the weather guidewas proposed and approved by the Forest andFire Meteorology Working Group, CanadianInteragency Forest Fire Centre, which waschaired at the time by Nick Nimchuk.

Gordon Miller, former director general, andBrenda Laishley, head of Publications, Northern

Forestry Centre, Canadian Forest Service,facilitated the publication process.

Alice Solyma, librarian, Pacific Forestry Centre,Canadian Forest Service, helpfully provided out-of-print versions of the World Meteorological

Organization standards.

We are indebted to Marty Alexander, seniorfire behavior research officer, Northern ForestryCentre, Canadian Forest Service, for his extensivesuggestions for updating the weather guide,and for contributing an appendix on latitudeconsiderations, which will allow the CanadianForest Fire Weather Index System to be adaptedfor use in other countries.

We sincerely thank the following reviewersfor their constructive contributions: Richard Carr,Roger Desjardins, Bill Droog, Mike Flannigan,

Jim Goosen, Ben Janz, Nathalie Lavoie, RobMcAlpine, and Mike Wotton.

THE AUTHORS

Bruce Lawson retired from the CanadianForest Service (CFS) in 1996 after 30 years asa forest fire research officer and head of thePacific and Yukon regional fire program. Hewas a member of the CFS Fire Danger Group,

contributing research and technology transferto the development of the Canadian ForestFire Weather Index System and the CanadianForest Fire Behavior Prediction System. In 1996,he began consulting in forest fire science andmanagement, mostly with Ember ResearchServices Ltd., undertaking projects in thedocumentation of wildfire behavior, developmentof fire management plans for parks and protectedareas, assessment of community fire risk, and

investigation of the cause and origin of fires forcourt cases.

Brad Armitage started working in fire researchwith the Pacific Forestry Centre of the CFS in

1989. While employed by the CFS, he workedon a number of fire research projects, includingignition probability modeling, FBP Systemtesting, prescribed burning, and site preparation.In 1994 he started Ember Research ServicesLtd. and began consulting in forest fire scienceand management for a variety of provincial andterritorial governments and for private industryclients.

THE PUBLICATION

Publication of this weather guide was fundedby the provinces of British Columbia and Albertaand by the Canadian Forest Service (CFS).This publication is a revision of the 1978 CFSpublication BC-X-177, titled Weather in theCanadian Forest Fire Danger Rating System A

User Guide to National Standards and Practices,by J.A. Turner and B.D. Lawson. Like itspredecessor, this weather guide is intended fornationwide use as a standard reference for firemanagers and researchers using the CanadianForest Fire Danger Rating System.


14/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM XII


15/87

INTRODUCTION

Weather accounts for all of the essentialinputs to the Canadian Forest Fire WeatherIndex (FWI) System, and these weather inputs,

together with outputs from the FWI System,are also required to calculate outputs from theCanadian Forest Fire Behavior Prediction (FBP)System. The FWI and FBP systems are the twoprincipal subsystems of the Canadian Forest FireDanger Rating System (CFFDRS) (Fig. 1).

The four weather elements that are measuredand used as inputs to the FWI and FBP systems

(rain accumulated over 24 h, temperature,relative humidity, and wind speed) are generallytaken daily at noon local standard time (LST)

or 1300 local daylight time (LDT). (The termnoon is used throughout this weather guide,even though most of Canada now implementsdaylight saving time over the entire fire season;therefore, 1300 LDT is generally an acceptableapproximation of solar noon.)

Fire

Weather

Index

System

Fire Occurrence

Prediction

System

Accessory

Fuel Moisture

System

Fire Behavior

Prediction

System

Risk

(lightning and

human-caused)

Weather Topography Fuels

CFFDRS

Figure 1. Structure of the Canadian Forest Fire Danger Rating System (CFFDRS) (adapted from Stocks et al.

[1989]).


16/87

WEATHER GUIDE FOR THE CANADIAN FOREST FIRE DANGER RATING SYSTEM 2

Fire

weather

observations

Fuel

moisturecodes

Fine Fuel

Moisture Code(FFMC)

Duf Moisture

Code(DMC)

Fire Weather

Index

(FWI)

Drought

Code(DC)

Yesterdays

FFMC

Yesterdays

DMC

Yesterdays

DC

Fire

behavior

indexes

Temperature

Relative humidity

Wind

Rain

Wind

Initial Spread

Index

(ISI)

Buildup

Index

(BUI)

Temperature

Relative humidity

Rain

Temperature

Rain

Figure 2. Structure of the Canadian Forest Fire Weather Index System (adapted from Canadian Forest Service [1984]).

The six standard components three fuelmoisture codes and three fire behavior indexes of the FWI System (shown in Fig. 2) providenumeric ratings of relative potential for wildlandfire. The FWI System refers primarily to astandard pine fuel type but is useful as a generalmeasure of forest fire danger in Canada (VanWagner 1987). The three fuel moisture codes

follow daily changes in the moisture content ofthree classes of forest fuel with different dryingrates. Each moisture code is calculated in twophases one for wetting by rain and one fordrying and is arranged so that higher valuesrepresent lower moisture contents and hencegreater flammability (Van Wagner 1987).

The Fine Fuel Moisture Code (FFMC) is anumeric rating of the moisture content of litterand other cured fine fuels. The FFMC is anindicator of the relative ease of ignition andflammability of fine fuels.

The Duff Moisture Code (DMC) is a numericrating of the moisture content of loosely compactedorganic (duff) layers of moderate depth. The DMCis an indicator of fuel consumption in moderateduff layers and medium-sized downed woodymaterial.

The Drought Code (DC) is a numeric rating ofthe moisture content of deep, compact organic

layers. The DC is an indicator of seasonaldrought effects on forest fuels and the amount ofsmoldering in deep duff layers and large logs.

Some physical properties of the forest floorlayers associated with the three fuel moisture

codes are summarized in Table 1, where thedrying rates of DMC and DC, represented bytime lag (i.e., time to lose 1 1/e where e isthe natural base of logarithms, which has thevalue of 2.7182818or about two-thirds of thefree moisture content above equilibrium), havebeen revised from those published earlier byVan Wagner (1987).


17/87


Two intermediate fire behavior indexesrepresent fire spread rate and amount ofavailable fuel. The Initial Spread Index (ISI)is a numeric rating of the expected rate of firespread, which combines the effects of wind andFFMC on rate of spread without the influence ofvariable quantities of fuel. The Buildup Index(BUI) is a numeric rating of the total amount offuel available for combustion, which combinesDMC and DC.

The final fire behavior index, the Fire WeatherIndex (FWI), combines ISI and BUI to representthe intensity of a spreading fire as energy outputrate per unit length of fire front. This numericrating of fire intensity is suitable as a generalindex of fire danger throughout the forestedareas of Canada.

One basic value of each FWI component iscalculated per day to represent fire danger

conditions during the midafternoon peak burningperiod, assuming a normal diurnal weatherpattern (Van Wagner 1987). For rainy days,calculation of the various fuel moisture codes hasbeen standardized by taking into account firstthe effect of the rain, and then the appropriatedegree of drying.

These six standard components of the FWISystem are predictors of daily fire potential

(Alexander and DeGroot 1988). Because onevalue per day is determined for each component,the FWI System does not indicate hour-by-hourchanges, nor does it account for variations infuel type from season to season or from placeto place. However, it does provide referencescales that permit comparisons of fire dangerwith other days and other locations. The FWISystem makes it possible to reconstruct past firedanger conditions if suitable historical weatherrecords are available. Thus, a fire climatologycan be developed for comparison with known fireactivity (Turner 1973; Stocks 1974; Kiil et al.1977; Harrington et al. 1983; Amiro et al 2004;Parisien et al. 2004; Girardin et al. 2006; Lavoieet al. 2007). The fuel moisture codes continueto be studied for correlation with observed fuelmoisture content of litter and forest floor stratawithin a wide range of ecosystems (Otway et al.2007) and on burned and unburned sites (Abbottet al. 2007a). Each fuel moisture code conveysdirect information about various aspects ofwildland fire potential. For example, fires are notlikely to spread in surface litter with an FFMC lessthan about 74, the duff layer does not contributeto frontal fire intensity until the DMC reaches 20,and ground or subsurface fire activity tends topersist at DC values greater than 400 (Stocks etal. 1989).

Table 1. Physical properties of forest floor layers associated with the fuel moisture codes of the Canadian Forest Fire WeatherIndex System

Fuelmoisturecode

Forestoorlayer

Nominaldepth(cm)

Nominalload

(kg/m2)

Bulkdensity(g/cm3)

Raincapacity(mm)

Saturatedmoisturecontent

(%)Standardtime laga

FFMCb Litter 1.2 0.25 0.021 0.62 250 2/3

DMCc Looselycompacted duff 7 5 0.071 15 300 15d

DCe Deep

compactedorganic layer 18 25 0.139 100 400 53d

aTime lags of the fuel moisture codes vary with weather conditions. Tabulated values represent standarddrying conditions (temperature 21.1 C, relative humidity 45%, wind speed 13 km/h, July) and were derivedby S.W. Taylor (Canadian Forest Service [CFS]) and veried by B.M. Wotton (CFS) and C.E. Van Wagner(CFS, retired).

bFFMC = Fine Fuel Moisture Code.cDMC = Duff Moisture Code.dDiffers from time lag presented in Van Wagner (1987), which is slightly in error.eDC = Drought Code.


18/87


The Daily Severity Rating (DSR) was describedby Van Wagner (1987) as an optional componentof the FWI System that is computed directly fromthe FWI. The DSR weights the FWI as it rises, ina manner deemed to reflect difficulty of controlin more direct proportion to the work required tosuppress a fire. The FWI itself is not consideredsuitable for averaging and should be used onlyas a simple daily value. Any averaging, whetherspatially over a number of stations on a givenday or at a single station over any period of time,is better accomplished through the DSR.

The DSR averaged over an entire fire seasonis termed the Seasonal Severity Rating (SSR),which can be used as an objective measure offire weather from season to season or of fireclimate from region to region.

The standard daily FFMC describes theafternoon state only (as forecast from noon

weather observations), and other means arerequired to describe fine fuel moisture at othertimes of the day. Van Wagner (1977) developedan hourly FFMC for which hourly weatherobservations are used to calculate an FFMC foreach hour around the clock. Similarly, Lawsonet al. (1996) presented look-up tables andcomputer coding for a diurnal FFMC calculatedfor each hour around the clock without the needfor hourly weather data. This diurnal FFMC wasan update of earlier tabular versions (Muraro etal. 1969; Van Wagner 1972; Alexander 1982b).Hourly and diurnal FFMC models are comparedin Appendix 2.

Currently, several Canadian fire managementagencies use the hourly FFMC, together withhourly wind speed and direction, as inputs to theFBP System (Forestry Canada Fire Danger Group1992), when quantitative estimates are requiredof variables such as head fire spread rate, fuelconsumption, fire intensity, and fire descriptionfor some 16 discrete fuel types. The FBP Systemis intended to supplement the experience andjudgment of operational fire managers. FBPSystem applications are more site specific andtime sensitive than FWI applications. Therefore,as noted by the Forestry Canada Fire Danger

Group (1992), the hourly FFMC computationalmethod is preferable for the prediction of firebehavior to both the standard daily FFMC, whichhas a standard diurnal curve embedded in it, andthe single diurnal curve in the table presented byVan Wagner (1972). In this way, hourly weathervariations, rather than average diurnal weathertrends, can be reflected in the fire behaviorpredictions.

The Accessory Fuel Moisture (AFM) Systemof the CFFDRS (Fig. 1) includes several fuel-specific moisture codes of the FWI System. Twoexamples illustrate the wide range of weather andfuel characteristics covered by such nonstandardmoisture codes. First, a sun-exposed fine fuelmoisture model was developed to represent the

moisture content of reindeer lichen (Pech 1989),for which the equilibrium moisture content withexposure to sun is 3%4% lower than for shadedlitter, the fuel represented by the FFMC. At theother end of the scale for exposure of fuel toweather elements is the sheltered duff moisturecode (Wotton et al. 2005), in which a modifiedDMC represents the moisture content of forestfloor fuels adjacent to tree boles, an area wherelightning strikes usually ignite the forest floor.These fuels are consistently lower in moisturecontent than less sheltered areas of the stand,because of greater interception of precipitationby tree crowns. Development of the AFM Systemis continuing (Alexander et al. 1996), without adefined end point or specified official content.

The Canadian Forest Fire OccurrencePrediction (FOP) System is envisioned as anational framework of both lightning- andhuman-caused fire components (Alexander etal. 1996). Although elements of an FOP Systemhave been developed using one or more FWISystem components (Anderson 2002; Wottonand Martell 2005), they have not yet beenimplemented on a national basis to predict thenumber of fires in specific areas.

Recent progress has been made on developingprobabilistic models of sustained flaming (Lawsonand Dalrymple 1998; Beverly and Wotton 2007)and smoldering ignition (Lawson et al. 1997;Anderson 2000; Otway et al. 2006).

Conceptually, the CFFDRS deals with theprediction of fire occurrence and behavior frompoint-source weather measurements (i.e., asingle station within a fire weather network) (Leeet al. 2002). The system does not account forspatial variation in weather elements betweenpoints of measurement. Models and othersystems external to the CFFDRS must be used

to handle such interpolation (see subsectionImplications of Fire Weather: Fire ManagementInformation Systems within the section FireWeather Forecasting). However, Lee et al. (2002)emphasized the inherent difficulty of obtainingsufficiently accurate and timely forecasts ofthe fire weather elements (most notably windspeed), especially for rugged terrain. Those


19/87


authors also noted the resulting limitations onany computerized decision-support systems thatdepend in whole or in part on the CFFDRS as ameans of predicting wildland fire occurrence andbehavior.

Computer-based fire management systems

have been used in Canada since the early1970s (Lee et al. 2002). In 1992, the CFSbegan investigating the use of geographicinformation systems (GIS) for constructingthese fire management information systems,which culminated in development of thespatial fire management system (sFMS).The fire-weather-related implications of this

technological advancement for the CFFDRS arediscussed later, under the heading Implicationsof Fire Weather: Fire Management InformationSystems.

The development of remote automaticweather stations (see the subsection entitled

Automatic Weather Stations within thesection Fire Weather Stations) and associatedcommunications technology in the 1980s and1990s permitted collection of weather datafrom isolated locations in almost real time on aprovincial and even a national basis (Taylor andAlexander 2006).

ELEMENTS OF FIRE WEATHER

The four weather elements needed to calculatethe six components of the FWI System are rain,temperature, relative humidity, and wind speed(Fig. 3). These elements influence the ease withwhich fires can be started, the rate of spread, andthe difficulty of controlling fires that are alreadyburning. Variation in day length throughout theseason affects both the DMC and the DC and isaccounted for by monthly adjustment in theirrespective daily drying factors. For these twoslow-reacting moisture codes, the amount ofmoisture lost daily by their representative fuelsis dependent as much on the time available as onthe noon atmospheric conditions. In contrast, themidafternoon moisture content of the fast-dryingfuels represented by the FFMC is less dependenton day length (Van Wagner 1987). The effect oflatitude on day length within the context of DMCand DC drying factor adjustments for countriesat various latitudes is discussed in Appendix 3.

When wind speed is determined, wind directionis also recorded, even though it is not requiredfor FWI System calculations. Wind direction is arequired input for calculations in the FBP System,is useful for interpolation of wind speed, and isimportant to those forecasting fire weather.

Temperature

The noon (dry-bulb) temperature (measuredin degrees Celsius) is required for the calculationof all three fuel moisture codes, FFMC, DMC, andDC.

Temperature, together with relative humidityand wind, affects the rate at which the FFMCrecovers after it has been reduced by rain.The recovery of the FFMC for three levels oftemperature, relative humidity, and wind isillustrated in Table 2.


20/87


Temperature

Relative humidity

Wind speed

Temperature

Relative humidity

Month

Temperature

Month

Rain?

No Yes

FFMC

Yesterday's

FFMC

Rain code

Rain?

Yes

Rain code

Drying factor

No

DMC

Yesterday's

DMC

Rain?

No

Drying factor

Rain code

DC

BUI

FWI

ISI

Yesterday's

DC

Yes

Wind speed

Rain Rain Rain

Figure 3. Schematic diagram for calculating the six standard components of the Canadian Forest Fire Weather Index System(adapted from Lawson

[1977]).FFMC = Fine Fuel Moisture Code, DMC = Duff Moisture Code, DC = Drought Code, ISI = Initial Spread

Index, BUI = Buildup Index, FWI = Fire Weather Index.

Table 2. Recovery of Fine Fuel Moisture Code (FFMC) after rain with three levels of temperature (temp.), relative humidity(RH), or wind speed (WS), with starting FFMC of 70

Days

sincerain

FFMC with variabletemperature, RH = 45%,

WS = 18 km/h

FFMC with variable relativehumidity, temp. = 20 C,

WS = 18 km/h

FFMC with variable windspeed, temp. = 20 C,

RH = 45%

10 C 20 C 30 C 65% 45% 25% 4 km/h18

km/h32

km/h

0 70 70 70 70 70 70 70 70 70

1 80 84 87 79 84 88 82 84 85

2 84 87a 89a 82 87a 91a 86 87a 87a

3 85a 85a 87a

aEquilibrium.


21/87


Temperature and relative humidity are bothrequired for calculating the drying phase of theDMC, whereas only temperature is needed tocalculate the drying phase of the DC. However,the drying factors for both DMC and DC aremodified by a day-length factor that varies bymonth.

Relative Humidity

The ability of the atmosphere to retainmoisture depends in large part on atmospheric

temperature. The higher the temperature, themore moisture the atmosphere can retain.Relative humidity (expressed as a percentage)is the fraction of moisture present in theatmosphere at a given temperature relative tothe total amount of moisture that the atmospherecould retain at that temperature. On a normalday when no significant moisture is added orremoved from the atmosphere, relative humidityvaries with temperature in a recognizable pattern(Fig. 4).

0100 0400 0800 1200 1600 2000 2400

0

2

4

6

8

10

12

14

16

18

20

22

24

Tem

perature(C)/andwindspeed(km/h

)

Relativehumidity(%)

0

10

20

30

40

50

60

70

80

90

100

Local standard time

Wind speed

Relative humidity

Temperature

Figure 4. Daily patterns of temperature, relative humidity, and wind speed in July for a typical continental station, Leighton Lake, B.C.

In the early days of organized forest firecontrol, the term fire weather meant relativehumidity, and this weather element is still usedtoday to quickly assess fire danger. However,the complexity of the problems that require fireweather inputs demands a more sophisticatedapproach. In particular, the FWI System requiresnoon relative humidity for determination of bothFFMC and DMC.

Relative humidity affects the day-to-day (orhour-to-hour, in the case of hourly FFMC) dryingrate of the FFMC in a nonlinear (logarithmic) way.This is discussed in more detail in the sectionSensitivity of CFFDRS Components to WeatherChanges. Relative humidity also affects theequilibrium moisture content (EMC), which is thelowest moisture content that a fuel will reach fora given combination of weather conditions. It is


22/87


useful to keep in mind that the EMC for the FFMCcovers a range that is greater than the range offlammability; for example, the EMC at relativehumidity of 100% is 35%, which is above theupper limit for fine fuel ignition of about 30%moisture content, whereas the EMC at relativehumidity of 10% is about 2%3% (Van Wagner1987).

By contrast, the DMC is based on anassumption of constant EMC (20%), i.e., doesnot vary with relative humidity. This assumptionmeans that regardless of how high the DMCclimbs, the lowest level of implied forest floormoisture is 20%. The daily drying factor that isadded to the DMC varies linearly with temperatureand relative humidity, but again, as with theFFMC, the relation between relative humidityand implied forest floor moisture content in theDMC is nonlinear (logarithmic). DMC is discussedin more detail in the subsection Duff Moisture

Code and Drought Code.

Wind

Wind (measured in kilometers per hour)influences the FWI System in two ways. Arelatively weak effect is felt in the daily changein the FFMC, for which wind speed chiefly affectsthe rate of recovery after rain. A much strongereffect is built into the ISI to reflect the jointinfluence of wind and moisture content of finesurface fuels on a fires rate of spread.

As a rule of thumb, the ISI doubles in value

for each increase of 14 km/h in wind speed, withFFMC held constant (Table 3). At the same time,with wind held constant, an increase of five toseven FFMC units is required to double the ISIunder moderate to severe conditions (Table 3).

Rain

The moisture content of forest fuels can beraised to 300% or more by contact with liquidwater, while a maximum fiber saturation value ofabout 30% for dead woody fuels in a saturatedatmosphere is possible (Schroeder and Buck

1970). Precipitation, usually in the form of rain,is the only factor that allows FFMC to fall below73. Expressed another way, rain is needed if finefuel moisture content is to exceed 31% (a valuederived from the following standard conversionformula: moisture content [%] = 147.2 [101 FFMC]/59.5 + FFMC; if FFMC = 73, moisturecontent = 31%). For further discussion on thederivation of this equation (FFMC/MC) see VanWagner (1987). Rain is also the only means bywhich DMC and DC can be reduced to valueslower than those recorded the previous day.

The total rainfall over 24 h (measured in

millimeters) must exceed certain thresholdamounts before it is considered to have anyeffect on the moisture content of the fuelsrepresented by the three fuel moisture codes.These threshold values are specific to each fuelmoisture code (Table 4). The effectiveness ofany given rainfall in reducing the value of eachmoisture code varies with the amount of therainfall and the value of the code before therain started. These variations in effectivenessare built into the FWI System to reflect what isknown about interception and rates of absorption(Table 5). Precipitation is measured in the open,but its effects are related to fuel moisture content

within forest stands.

From time to time during the fire season,precipitation may occur as hail or snow. In manycases precipitation that falls in this form willmelt in the interval between observations, sothe equivalent depth of water is entered into the

weather record as if it had been rain.

If the snow (or hail) remains on the groundat observation time, the calculation of thethree moisture codes continues, using thewater equivalent of snow that has fallen sincethe previous observation (where 1 cm of snow= 1 mm of rain). However, the ISI and FWIboth have the value zero under these conditionsand retain this value until the snow or hail hasmelted.

Table 3. Effect of wind on Initial Spread Index (ISI)

ISI at various wind speeds

FFMCa5

km/h19

km/h33

km/h47

km/h

77 1 2 5 9

80 2 3 6 12

83 2 4 9 17

86 3 6 13 26

89 5 10 20 40

92 7 15 30 61

95 11 23 46 93

98 17 34 68 138

aFFMC = Fine Fuel Moisture Code.

Table 4. Threshold rain values for fuel moisture codes

Fuel moisture code 24h rain (mm)

Fine Fuel Moisture Code > 0.5

Duff Moisture Code > 1.5

Drought Code > 2.8


23/87


Table 5. Effects of rain on the fuel moisture codes

Todays rain code(% reduction of yesterdays value)

24-h rain(mm)

FFMCa

(yesterday = 90)DMCb

(yesterday = 30)DCc

(yesterday = 100)

0.6d 86 (4)1.6e 68 (24) 29 (3)

2.9f 48 (47) 25 (17) 195 (3)

5.0 33 (63) 20 (33) 192 (4)

10.0 19 (79) 15 (50) 178 (11)

20.0 14 (84) 12 (60) 155 (23)

40.0 10 (89) 10 (67) 114 (43)

aFFMC = Fine Fuel Moisture Code.bDMC = Duff Moisture Code.cDC = Drought Code.dJust exceeds threshold value for FFMC.eJust exceeds threshold value for DMC.fJust exceeds threshold value for DC.

Supplementary Weather Elements

Fire weather observation programs are builtaround the needs and standards of the FWISystem, but additional information that is notpart of the FWI System is often required forfire management purposes and for fire weatherforecasting. Standards for such supplementaryinformation are specified by individual forestmanagement agencies to meet regionalrequirements.

The following are examples of the additionalinformation that may be required:

Basic fire weather elements: Observationsof basic fire weather elements, includingextremes of relative humidity andtemperature, may be recorded at timesother than noon.

Wind direction: Fire weather forecastersuse wind direction to relate local windpatterns to broad-scale wind flow andtopographic features. Wind direction is arequired input for the FBP System.

Lightning occurrence: Most of the forestedarea of Canada is covered by automatic

networks for lightning detection andlocation, and the data are readily accessibleto fire weather forecasters.

Cloud conditions: Information about thedevelopment and movement of lightning-producing clouds (Mullock 1982) and ceilingheights may be required for deployment ofaircraft and for fire weather forecasting.

Dew: The effect of dew on fuel moistureis generally limited and dissipates by noon

(see subsection entitled Dew and Frostwithin the section Weather Not DirectlyObservable).

Upper atmosphere profiles: See subsectionentitled Low Level Jet within the sectionWeather Not Directly Observable.

Solar radiation: Some automaticweather stations record the duration ofbright sunshine (defined by the WorldMeteorological Organization [WMO 2006]as direct solar irradiance > 120 W/m2).The US National Fire Danger Rating System(NFDRS) calculates 10-h fuel moisture

on the basis of solar radiation measuredhourly. Sixty 1-min samples are averagedover a 1-h period before data transmission(NWCG 2005).

Fuel moisture: Some automatic weatherstations have sensors that measurevariables related to the fuel moisturecontent of wooden dowels (see subsectionAutomatic Weather Stations within thesection Fire Weather Stations). However,the US NFDRS recommends that directmeasurements of fuel moisture sticks beused in calculations for that system (NWCG2005).

Snow depth: Snow depth is used forstart-up and shut-down of FWI Systemcalculations at the beginning and end ofeach fire season.

Atmospheric pressure: Various weathermodels require input of atmosphericpressure, including adjustment oftemperature and humidity to reflectelevation.


24/87


WEATHER OBSERVATION PRACTICES

Weather observation practices have beencarefully specified and must be followed asclosely as possible to ensure the effectivenessof management decisions that are based onthe results. Such standards are essential forrelatively permanent fire weather stations thatform a regional network. It may be necessary torelax the standards for short-term stations setup to provide on-the-spot weather reporting forspecific purposes.

Time of Observations

Basic Observation Time

Weather readings are taken at noon,1200 LST or 1300 LDT when and where the

latter is in effect. Weather recorders should beset to the exact hour, without correcting for sunnoon differences at individual weather stationlocations.

Noon was chosen as the basic observation timeto ensure that weather readings are taken lateenough in the day to indicate conditions duringthe period of afternoon peak fire activity but early

enough that codes, indices, and forecasts will beavailable for planning and operational purposes.Weather observations should be taken within

15 min of the specified time. If this specificationis followed, temperature, relative humidity, andwind are unlikely to be sufficiently in error toreduce significantly the accuracy of the FWISystem calculations (Table 6).

Table 6 shows that around noon at typicalCanadian weather stations, the temperature isincreasing by, on average, less than 1 C per hour,the relative humidity is dropping by less than 4%per hour, and wind speed is increasing by about0.5 km/h per hour. Table 6 is based on data forall days between May and October, includingcloudy, rainy, and clear days. Somewhat larger

changes in temperature, relative humidity, andwind speed, as much as 60% greater than thehourly rates of change shown in Table 6, can beexpected on clear days, and of course the ratesof change for any of these weather elementson any particular day could greatly exceed theaverage shown.

Table 6. Average hourly rates of change of temperature, relative humidity (RH), and wind speed at noon local standard timefor select stations

Station Latitude Longitude

Change intemperature

(C/h)Change inRH (%/h)

Change in windspeed

(km/h perhour)

Gander, NL 48.96 54.61 0.55 2.6 0.66

Chatham, NB 47.00 65.45 0.74 3.5 0.69

Bagotville, QC 48.30 71.00 0.70 2.8 0.61

Kapuskasing, ON 49.42 82.42 0.75 2.8 0.50

The Pas, MB 53.81 101.24 0.64 3.2 0.42Fort McMurray, AB 56.72 111.40 0.89 4.2 0.60

Prince George, BC 53.91 122.78 0.77 3.0 0.47

Port Hardy, BC 50.72 127.47 0.52 2.2 0.58

Whitehorse, YT 60.73 135.08 0.66 2.4 0.26

Average 0.69 3.0 0.53

Source: Turner and Lawson (1978); reprinted with permission of Pacic Forestry Centre.


25/87


Deviations from Basic Observation Time

Deviations from the basic observation timeof 1200 LST may be specified at the regionallevel for stations that are close to time zoneboundaries, are at high latitudes, or both.

Van Wagner (1987) quantified the effectsof latitude on calculated values of FFMC, DMC,and DC. Recommended day-length factors forDMC and seasonal adjustments for DC for moresoutherly latitudes than Canada are discussedin Appendix 3. However, Van Wagner (1987)noted that it was fair to question whether theFWI System should take into account the effectof latitude within Canada on FFMC, since thiscode incorporates no allowance for day length.Van Wagners (1987) comparisons of standarddaily FFMC against hourly FFMC calculated fromweather observations taken at 1600 (calculatedas FFMC at 1600 minus standard FFMC) produced

discrepancies ranging from 1.1 for stations at48 N to 2.5 for stations at 66 N.

Van Wagner proposed that the basic observa-tion time could be delayed progressively fromnoon at lower latitudes to about 1400 at higherlatitudes to eliminate much of the discrepancy instandard daily FFMC and to accurately accountfor daily peak fire danger conditions at higherlatitudes. The FFMC discrepancy occurs becausethe time of maximum temperature and minimumrelative humidity in high summer (the several-week period around summer solstice) isprogressively later as latitude increases. Standard

daily FFMC has a built-in forecast mechanismthat assumes that daily peak conditions occurat 1600 h LST, but north of 60 latitude, thedaily peak of temperature and trough of relativehumidity tends to be delayed until 1800 LST oreven later, and tends to persist for longer thanthe 1 h or so that is typical for midlatitudes,particularly around high summer.

Kiil and Quintilio (1975) compared Junediurnal relative humidity cycles for a 10-yearperiod and found that maximum humidityovernight was significantly lower in the NorthwestTerritories than at lower latitudes. As expected,maximum overnight temperatures were higherat high latitudes, reflecting the absence of thelong, cool, moist nights that are common in thesouth. Both Kiil and Quintilio (1975) and Wardand Mawdsley (2000) noted that northern firesare known to burn with high intensity aroundthe clock, presumably because of the combinedeffects of lack of recovery of fuel moisture andambient weather conditions conducive to rapidfire spread.

According to Van Wagner (1987), theabove discrepancies were not judged seriousenough to warrant an official recommendationthat procedures be revised to adjust the basicobservation time for latitude, since the standarddaily FFMC measures peak flammabilityreasonably well at all latitudes.

However, one other factor contributes to theunderprediction by the FFMC of daily peak firedanger conditions north of 60 latitude duringhigh summer. In the Yukon Territory and theNorthwest Territories, the basic observation timeof 1200 LST can deviate significantly from sunnoon (Paul 1974) because of regional adoptionof politically rather than geographically basedtime zones and the longitudinal location of fireweather stations within the time zones.

Inuvik, N.W.T., offers a good example.Located geographically near the Arctic coast and

just east of the Yukons eastern boundary, Inuviklies two universal time zones west (i.e., behind)the political time zone of mountain standard time(MST) in which it functions. Universal time (UT)is defined as the time of the zone centred onthe zero (prime) meridian through Greenwich,U.K., with each of the other time zones a definitenumber of hours ahead or behind UT to a total of12 hours (Dominion Bureau of Statistics 1971).As such, 2000 Greenwich Mean Time (GMT) or20 z corresponds to mountain standard time of1300 (i.e., z minus 7 h). Even the Pacific standardtime (PST) zone (z minus 8 h) in which BritishColumbia and Yukon function is geographically

too far to the east, on the basis of universaltime (UT) zones, to capture Inuvik. Inuvik, aswell as most of the Yukon, lie geographically ina z minus 9 h time zone that is simply not usedin Canada.

The UT zones are the idealized 24 time zonesthat resulted when standard time was establishedat a world conference held in Washington, D.C.,in 1884. Ideally, each time zone covers an areadefined by two meridians of longitude 15 apart.However, in practice, because of political andgeographic considerations, the boundaries ofindividual time zones are extremely irregular,

exemplified by the pronounced extent to whichthe Yukon and Northwest Territories fall withingeographically incorrect time zones. Theterritory of Nunavut is unaffected, as westernNunavut operates in the mountain time zoneand the central portion of the territory operatesin the central time zone, essentially a politicalmatch to the correct geographic time zones.


26/87


The following compromise is recommendedto the problem of Yukon and the NorthwestTerritories operating one and two time zones,respectively, ahead of what geography alonewould dictate: for these two territories, basicobservation time for fire weather should bemoved ahead by 1 h, to 1300 PST for Yukonand to 1300 MST for the Northwest Territories.Although communities in the northwest portion ofthe Northwest Territories (such as Inuvik) couldjustify a 2-h adjustment of basic observation time,it is assumed that such a change would introduceadditional problems, such as marked delays inthe availability of fire danger information andsignificant changes to long-standing calibrationsof FWI System components.

Recording Practices

The standard noon weather observations(rain, temperature, relative humidity, and wind)

and the FWI System calculations should berecorded daily on a permanent monthly form.One such form for manually recording weatherobservations and table-based FWI Systemcalculations is provided on the inside back cover ofthe FWI System tables (Canadian Forest Service1984). However, that form does not containcolumns for remarks or other weather variables,such as cloud cover, visibility, and maximumand minimum temperatures. These data, whichare of value to fire weather forecasters, maybe added to forms and the relevant collectionprocedures may be specified by regional fireweather authorities.

Precision Standards andAccuracy of Measurement

The terms accuracy, precision, and sensitivityare all used from time to time in connection withfire weather measurements and danger ratingscales. A few words of explanation may clear upconfusion among them.

The accuracy of a measurement is related tothe instrument or technique of measurement.When an instrument is described as having anaccuracy of 5 units, this generally means that

a series of measurements of some constantproperty made with the instrument were mostly(95% of the time) within 5 units of the correctvalue.

Precision is concerned with the size of theunit (or the number of decimal places) usedin taking and recording a given measurement.To say that a given length is 6 m implies that

the true measurement lies somewhere between5.5 and 6.5 m. The same length expressed as600 cm implies a precision of 1 cm (i.e., themeasurement fell between 599.5 and 600.5).Precision may be expressed as a fraction (e.g.,0.5 C) or as a round number (5%).

In general, to take full advantage of theaccuracy of a particular measuring system, theprecision is specified to the next whole unit belowthe range of accuracy of the equipment. Forexample, relative humidity is normally measuredand recorded to the nearest whole percent, eventhough the accuracy of the equipment may be 5%.

Sensitivity relates to the amount of changein a measurement or a derived index that isproduced by a given change in the property beingmeasured or in one of the component factors.As such, sensitivity is a relative property. For

example, some types of relative humidity sensorare more sensitive to changes in relative humidityat lower values than they are near saturation.

Precision standards for recording, and theaccuracy of weather instruments required formeasuring temperature, relative humidity, windand rain as inputs to the CFFDRS are specifiedin the subsections immediately following, whilesensitivity of calculated CFFDRS components toweather changes is discussed in its own section.

Temperature

The FWI System tables (Canadian Forest

Service 1984) offer the following instruction formeasuring temperature:

Observe the dry-bulb and wet-bulbtemperatures, using ventilated thermometers,and record to at least the nearest one-half(0.5) degree Celsius. The preferred instrumentis an electric fan psychrometer housed in aStevenson screen.

These instructions reflect the normal precisionrequired for dry-bulb temperature observationsfor FWI System calculations, which is the nearest0.5 C. The accuracy of the thermometersor temperature sensors should therefore bebetter than this (i.e., they should be accurate to 0.1 C).

Wet-bulb temperatures are measured tothe same precision, with thermometers havingcomparable accuracy and response time to thoseused for dry-bulb temperatures.


27/87


A precision of 0.5 C for the wet- and dry-bulb temperatures leads to a precision of 1 Cin the wet-bulb depression, such that calculatedvalues of relative humidity from Table 10 in thestandard FWI System tables (Canadian ForestService 1984) are significant only to the nearest5%.

Relative Humidity

The FWI System tables (Canadian ForestService 1984) give the following instructions formeasuring relative humidity:

Determine the relative humidity from dry-bulband wet-bulb temperatures and record to thenearest percent. Three RH tables are includedin this publication for use within the followingelevation ranges:

RH Table Elevation (m)

10A 0 to 305

10B 306 to 760

10C 761 and higher

Use the table appropriate for the stationelevation.

The general accuracy of RH values determinedfrom ventilated wet- and dry-bulb temperatureswill be well within the requirements for fireweather, provided the thermometers are accurateto the limits specified in the previous section.

Relative humidity values taken from recordinghygrographs are generally significant only tothe nearest 5%, provided conditions are notchanging rapidly. As for calculated values ofrelative humidity, hygrograph readings may berecorded to the nearest percent.

Electric fan psychrometers, either withina Stevenson screen or as portable models,should be run for at least 20 s before anymeasurements are taken, to be sure that thefull wet-bulb depression has been reached. Slingpsychrometers must be twirled for at least 20 s toensure that constant values have been reachedbefore temperatures are read (first the wet-bulb

and then the dry-bulb temperatures). Care mustbe taken to shield the unit from direct sun whentwirling and taking readings.

Wind

The FWI System Tables (Canadian ForestService 1984) give the following instructions formeasuring wind:

Measure and average wind speed over at leasta 10-minute period and record to the nearestwhole km/h. Wind speed should be measuredwith a cup-type counting anemometer, not withan instantaneous wind indicator. Preferably, theanemometer should be located in the open at aheight of 10 m above the ground.

Interpretation:

As noted in these instructions, wind isusually measured by a wind vane and cup ora propeller anemometer. Wind speed shouldbe recorded as an average of the preceding10 min of observations. Modern wind-measuringsystems contain not only the sensors but alsoa processing and recording system that takescare of the averaging and which may alsocompute standard deviations, extremes, andgustiness. Peak gust is the maximum observedwind speed over a specified time interval (e.g.,the last full hour in an hourly weather reportingsystem). Anemometers should have a responselength of less than 5 m, which is a measure ofthe equipments responsiveness to a change inwind speed. The 10-min averages of wind speedshould be based on 0.25-s samples.

As mentioned in the subsection Supplemen-tary Weather Elements in the section Elementsof Fire Weather, wind direction is a required in-put to the FBP System; as such, it has becomea standard weather observation at fire weatherstations. Wind direction should be reported indegrees, to the nearest full degree. As with windspeed, recorded wind direction should representa 10-min average. Wind direction is defined asthe direction from which the wind blows, mea-sured clockwise from geographic north (i.e., truenorth). Wind direction should be measured withan accuracy of 5.

Rain

The FWI System Tables (Canadian ForestService 1984) provide the following instructionsfor measuring rain:

Measure the rain accumulated in the 24-hourperiod from noon to noon, and record to thenearest one-fifth (0.2) mm. Locate the rain

gauge on the ground in the open.

In the case of snow, measure the averagedepth in cm and record the water equivalent asthe same number of mm. For example 2.4 cmof snow is reported as 2.4 mm of rain.

These instructions specify that rainfall is tobe measured with a precision of at least 0.2 mm.


28/87


In practice, even though the accuracy of tipping-bucket rain gauges or manually interpreted rain-gauge graduates is 0.2 mm, rainfall is recorded tothe nearest 0.1 mm (i.e., the next smaller unit ofprecision below the accuracy of the equipment).

In the case of snow or hail and when it isreasonable to suppose that the amount collectedin the rain gauge is an accurate sample, theactual depth of the meltwater is used. Forheavier snowfalls or hailstorms, when the gaugetypically cannot catch a representative sample,an average value of the depth of snow or hail onthe ground should be measured with a ruler to atleast the nearest 0.2 cm.

Sudden Weather Changesduring the Afternoon

Account must be taken of the sudden weather

changes that frequently occur on summerafternoons; otherwise, the abrupt changes inrelative humidity or wind or the occurrence ofafternoon showers after calculation of the FWISystem components will result in misleadinginformation. Up to 1600 LDT, it is acceptableto create a revised index for the day to reflectmore accurately the new weather regime. Inthis situation, a new set of weather observationsshould be obtained and the FWI Systemcomponents recalculated. For the purposes of thissupplementary calculation, the values calculatedat noon are ignored, and new values of FFMC,DMC, and DC are calculated, using the previous

days values as the starting fuel moisture codes.The amounts of rainfall used in the calculationsshould include any rain that has fallen sincenoon. This rainfall must also be included in the24-h amount at the next regular observation,but it is the noon fuel moisture codes that arecarried over to the next days calculations.

Extended periods of fog or low cloud arereflected in the index calculations only by theassociated high relative humidity and lowtemperature at observation time. As long asthe fog is present, the moisture codes may notfully represent the true moisture content of the

fuel complex, but after one full drying period,the moisture codes will be correct. If fog isconsistently present at the noon observationtime but clears within an hour or so, follow theprocedure outlined above, but use observationstaken after the fog has cleared.

For recording purposes, the value at noonremains the standard observation for the day.

This will normally provide the values from whichthe next days codes are calculated.

One basic assumption in the development ofthe FWI System is that the component weatherelements follow a more or less typical diurnalpattern, at least from late morning through late

afternoon (Fig. 4). However, in some locations,the regular pattern is distinctly different from thenorm. Stations subject to strong sea breezes orvalley winds, which pick up after noon, presentspecial problems. Valley bottoms or coastalstrips subject to morning fog that persists untilnoon are best handled by taking additionalobservations.

If hourly weather observations are available,it is possible to calculate FFMC for every houraround the clock using a computer program suchas that described by Van Wagner (1977). HourlyFFMC is recommended to establish diurnal

patterns of fire danger for unusual situationscreated by latitude, elevation, coastal or valleywinds, or other factors.

Seasonal Start-up andShut-down of Calculations

The FWI System tables (Canadian ForestService 1984) provide the following guidanceon starting seasonal recording of weatherelements:

Start the daily record as soon as there ismeasurable fire danger in the spring. The exact

date and starting values of FFMC, DMC and DCwill normally be provided by the appropriatefire weather authorities. In the absence of suchdirection, choose the starting date according tothe following criteria:

(a) In regions normally covered by snow duringthe winter, begin calculations on the third dayafter snow has essentially left the area to whichthe fire danger rating applies.

(b) In regions where snow cover is not asignificant feature, begin calculations on thethird successive day that noon temperatures of12 C or higher have been recorded.

In either case, use the following startingvalues:

FFMC 85; DMC 6; DC 15

These values should not be applied to late-starting stations. Contact the fire weatherauthority for instructions.


29/87


These instructions take into account the factthat data may be required for supplementarystations that cannot be put into operation atthe beginning of the season. In such cases,the regional fire weather authority should becontacted for an estimate of the start-up codevalues to be used, particularly for DMC and DC.An incorrect guess for the start-up FFMC willcorrect itself after about three days of recording.Do not assume automatically that the DMC andDC for a particular location will begin at standardstart-up values. For late-starting stations, thestart-up values of these two codes will generallyuse the current values of DMC and DC from oneor more nearby representative stations.

For stations starting up at the beginning ofthe fire season or later in the spring or summer,the DC value may have to be adjusted fordeficiency in precipitation over the winter. Theprocedure is complicated (see subsection entitled

Overwintering the Drought Code within thesection Drought), but the regional fire weatherauthority will generally be able to provide thenecessary over-winter adjustment for the DCstarting value.

The FWI System tables (Canadian ForestService 1984) also provide guidance on closingdown weather element recording for theseason:

Closing dates for fire danger calculations willnormally be supplied by regional fire weatherauthorities. In the absence of such direction,

it is recommended that observations andcalculations be continued until snow coversthe ground. Otherwise, the tables provide forthe calculation of the components of the FWISystem until the end of November.

These instructions for closing downobservations do not discuss a situation that mayoccur in the northern hemisphere, in which snowcover is absent after November 30 and noontemperatures in December remain above 12 C;under these conditions and in the absence ofrain, active drying of fuel may still be occurring.In this case, daily observations and calculationsshould continue until snow covers the ground or

noon temperatures drop below 12 C for threeconsecutive days.

Missing Observations

The FWI System tables (Canadian ForestService 1984) include the following informationabout dealing with missing observations:

The FWI System requires an unbroken dailyweather record. Days when observations aremissed cannot be ignored; the best possibleestimate of the missing weather observationsis necessary to preserve the continuity of thefuel moisture codes. By one means or another,therefore, blank spaces in the daily weather

record must be completed.

In the case of days when observations aremissed, contact the fire weather authoritiesfor instructions. In the absence of direction,complete the daily record by one of thefollowing methods, and circle the estimatedweather observation(s) on the record form.

(a) use values from recording instruments ifavailable on site;

(b) use average of values from one or morenearby similar stations;

(c) use average of values from day before andday after; or

(d) estimate values from knowledge of recentweather pattern.

These instructions apply to days whenobservations are unavoidably missed becauseof equipment breakdown or for some otherreason. Given that gaps in the record mean aloss of accuracy in the calculation of FWI Systemcomponents, these gaps should be minimized.The following procedure represents a consistentmethod of using available information to minimizethe errors caused by missing observations.

Measure total rainfall on the day after the1.

day (or period) of missed observations,and do your best to assign reasonableportions of that total to each day forwhich observations are missing (includingthe day of measurement). Check therecord for relative humidity (using thehygrothermograph chart or the hourlyrelative humidity record, if available) tohelp estimate the timing of rainfall andthus to determine if all of the recordedrainfall fell on one day.

Estimate the noon (or 1300 LDT) relative2.humidity and temperature from the

hygrothermograph chart, if available.

Assume the wind to be in the 413 km/h3.class for the FFMC calculation, unless youhave good reason to suspect that it shouldbe in one of the other classes.

After making the necessary estimates of4.rainfall, temperature, relative humidity,


30/87


and wind speed, calculate the codes andindices for each missing day as you wouldhave done if the observations had beenobtained in the regular way.

The following points should be rememberedwhen missing data are generated in this way:

A common error is to treat days withmissing data as if they did not exist, usingthe moisture code values for the last daybefore the gap as the starting point forcalculations on the next day of observations.This results in misleading (either low orhigh) values of DMC and DC. To avoidthis problem, separate calculations mustalways be performed for each missing day,based on the best possible estimate foreach required weather element.

If the procedures described above are notfeasible, because a hygrothermographis lacking or because the instrumenthas had a breakdown, try to get missingobservations from the nearest weatherstation or, even better, average the valuesfrom several stations.

Estimated values for wind and distributionof rain are usually adequate for thebookkeeping required to keep track of themoisture codes. However, the values ofISI and FWI calculated for those days maybe subject to large errors and should betreated with caution.

Effect of Surrounding Terrainon Measured Wind Speed

Although applicable to the midafternoon peakfire danger period, the weather elements usedto forecast the FWI System components are fornoon LST. Of the four key weather elements(rain, temperature, relative humidity, and wind),wind speed is the most difficult to forecast.Moreover, for forestry purposes, the forecastedwind velocity used for predicting fire dangeris necessarily lower than what is prepared forpublic forecasts, as explained below.

The roughness of ground and vegetation

surfaces affects wind speed, turbulence, andgustiness to a height of 600 m or so above meanground level, depending on atmospheric stability.Wind at the top of this friction layer is called thegradient or free-air wind.

A typical comparison of how surface roughnessreduces wind speed near the earths surfacewould involve observations from a 10-m mastlocated in an open grassy field and observationsfrom an opening surrounded by a forest of 15-mpines. The anemometer in the open field isexpected to record a wind speed of about 60%of the gradient wind, whereas only about 36%of the gradient wind speed would be measuredin the opening surrounded by pine stands. Citiesand urban areas in general are associated with aneven greater reduction in wind speed. Because ofthe scale of roughness of houses and commercialbuildings, a standard anemometer will show onlyabout 23% of the gradient wind if located in anopening surrounded by such structures.

For a weather station clearing that issurrounded by a 15-m pine stand, these relationsmean that the measured wind speed would beonly about 60% of that measured 10 m above a

large grassy field, assuming the same pressuregradient. Similarly, an anemometer located inan opening in an urban setting would measureonly about 40% of the wind speed measuredover open fields. These relations are shown inFigure 5 and are illustrated by the followingexample (from Turner and Lawson 1978):

Gradient wind at 600 m = 64 km/h

Wind at 10 m above extensive open grassland(0.60 64 km/h) = 38 km/h

Wind at 10 m in opening surrounded by15-m pine stand(0.36 64 km/h or 0.60 38 km/h) = 23 km/h

Wind at 10 m in opening surrounded by citybuildings(0.23 64 km/h or 0.40 38 km/h) = 15 km/h

Although the percentage reductions in windspeed owing to surface roughness are subjectto wide variations, they can be regarded astypical. Many airport locations give similar windspeed ratios compared to gradient winds as openfields compared to openings surrounded by pinestands, as given here (Simard 1971). A generallyacceptable rule of thumb for calculations in

the FWI System is to multiply the wind speedmeasured at an airport by 60% so that theyare comparable to winds measured in a forestopening.


31/87


Airport wind (A)

A = 0.6G

Forest wind (F)

F = 0.36G or 0.6A

Urban wind (U)

U = 0.23G or 0.4A

A

B

C

Figure 5. Reduction of surface wind speeds according to roughness of surrounding terrain. (A) Airport wind in relatively smooth open grassland. (B) Typical

forest wind, where surrounding timber slows the wind and creates turbulence. (C) Wind in a city opening, which is further reduced by surface roughness.

G = gradient wind (adapted from Turner and Lawson [1978].)


32/87


FIRE WEATHER STATIONS

Location Standards

In general, the standards for Canadianfire weather stations conform with those

recommended by the WMO for agrometeorologicobservations in forest areas (WMO 1968). Thestandards are designed to give representativevalues of the various weather elements;unfortunately, however, it is not usually possibleto meet all of these standards in practice.Nonetheless, every effort should be made to doso, since any major deviation can reduce theaccuracy of the FWI System components.

Ideally, the location of the fire weather stationshould have the following characteristics:

representative of the general area of

concern with respect to topography,vegetative cover, and local weatherpatterns, with avoidance of shelteredvalleys and exposed peaks and ridgetops, a preference for level or nearly levelground (Fig. 6) (or, if slopes must beused, avoidance of shaded and east-facingslopes), and avoidance of concave (dish-shaped) landforms;

at the center of a forest clearing withdiameter no less than 10 times the heightof the surrounding timber;

no closer than 100 m to any major source

of moisture, such as a lake, stream,swamp, or irrigated area;

no closer than 10 m (or, in the case ofbuildings, no closer than a distance equalto twice the height of the building) fromlarge reflecting or radiating surfaces,

such as metal or white-painted buildings,black-topped or graveled parking lots, rockoutcrops, and recently burned areas;

no closer than a distance equal to 1.5times the height of the obstruction fromany large building, tree, or vegetation;

no closer than 5 m from any road; and

at least 50 m away from excessively dustyareas (which can usually be avoided bychecking dust accumulation on nearbyvegetation).

If the prevailing wind direction during fairweather is known for the area, the station shouldbe located on the windward side of any sourcesof moisture, reflection, radiation, or dust.

It is good practice to arrange the instrumentsin a fenced enclosure at least 7 m 7 m. The typeof fencing is subject to regional specificationsbut should be of wire or open-pole constructionsuitable for safeguarding the equipment andnot more than 1.2 m high. The ground areainside the fence should consist of mown grass orcropped natural vegetation. When located in alogged area, the enclosure should be cleared oflogs and branches.

Figure 6. Fire weather station in large clearing on open level ground. Photo courtesy of Ember Research Services Ltd.


33/87


Instrumentation

This weather guide does not describethe maintenance of specific meteorologicalinstruments used for fire weather observations.Such specifications, which are too detailed for a

users guide of this type, are available in otherpublications (British Columbia Forest Service1969; Ontario Ministry of Natural Resources1974; Environment Canada 1977; Finklin andFischer 1990; WMO 2006); guidance may also beavailable from regional fire weather authorities.

Instrument Shelters

Thermometers and recording instrumentssuch as hygrothermographs must be housed ina white-painted, wooden Stevenson-type screenwith double-louvered sides and double roof(Fig. 7). The screen should be solidly mounted,

with the floor 115 cm above ground level andthe door opening to the north in the northernhemisphere or to the south in the southernhemisphere. The shelter should be mounted ona rigid but open framework of posts, not on asolid base such as a stump.

Figure 7. Stevenson screen, on an open framework stand that is staked to the ground. The floor of the unit is 115 cm

above ground. The white-painted double-louvered wooden door faces north on this unit, which is located in the

northern hemisphere. Photo courtesy of British Columbia Ministry of Forests and Range, Protection Branch.

Instrument screens should be large enough toproperly house the equipment they are designedto shelter. Auxiliary equipment not requiringthis kind of screen (e.g., data loggers) should

be housed in a separate box. It is particularlyimportant that screens be kept painted and ingood repair. In particular, they must be kept freeof dust and dirt, both inside and outside.

Although it is possible to specify shortresponse times for the thermometers andelectronic temperature and relative humiditysensors used in automatic weather stations,

these are usually overridden by the responsetime of the instrument shelter. Typical responsetimes are about 10 minutes for shelters in forestclearings, which is adequate for fire weather

purposes.

Small, white, louvered radiation shields forthe temperature and relative humidity sensorsused in automatic weather stations generallyhave response times of a few minutes, suitedto the shorter response times of electroniccapacitance-type relative humidity sensors.


34/87


Equipment for Measuring Temperature andRelative Humidity

The dry- and wet-bulb thermometers usedfor determining relative humidity should have anaccuracy of 0.1 C. They must be adequately

ventilated, preferably by a motor-driven fanwith the capacity to deliver air past the twothermometer bulbs with a velocity of at least3 m/s. One type of an electric fan psychrometeris shown in Figure 8.

Figure 8. Portable electric fan psychrometer, with battery-powered fan and wet-and dry-bulb thermometers.

Photo courtesy of British Columbia Ministry of Forests and Range, Protection Branch.

Dry- and wet-bulb readings may be taken witha good portable psychrometer with a battery-operated fan or with a sling psychrometer largeenough to provide the necessary precision. Thewicking for the wet-bulb thermometer must beclean and should be replaced several times aseason. Clean, mineral-free water or distilledwater should be used to wet the wick.

The hygrothermograph is an instrumentthat records temperature and relative humidityon the same chart (Fig. 9). The thermographcomponent should be capable of an accuracy of

0.5 C and should have a time constant on theorder of several minutes or less. The hygrographcomponent should be accurate to less than 5%

under steady conditions, where the relativehumidity is not changing rapidly. The hairs shouldbe arranged to have a short response time atnormal operating temperatures. The instrumentshould not have any significant temperaturecoefficient or, if it does have one, the correctionfactor should be known and should be applied tothe readings.

In practice, hygrothermographs may be quiteinaccurate, especially at low relative humidity,where accuracy is critical. Therefore, it is usefulto compare hygrothermograph chart readings

with relative humidity determined by a sling orelectric fan psychrometer.


35/87


Equipment for Measuring Wind

To meet the WMO standards on which theFWI System is based, the anemometer shouldbe mounted on a substantial, well-guyed mast,with provision for climbing with safety or forlowering the anemometer head when servicingis required. Provision for lightning protection ishighly recommended.

The three-cup anemometer, with itsruggedness and reliability, is well suited to firedanger measurements. It is a simple matter toelectronically count the number of meters of windthat pass the cups in the basic 10-min period.

Figure

Documents

2008 CFFDRS Weather Guide.pdf