EUFIRELAB

EVR1-CT-2002-40028

D-03-09

http://eufirelab.org

EUFIRELAB:
Euro-Mediterranean Wildland Fire Laboratory,
a “wall-less” Laboratory
for Wildland Fire Sciences and Technologies
in the Euro-Mediterranean Region

Deliverable D-03-09

Behaviour Modelling of Wildland Fires:

Final version of the State of the Art

Coordinated by Albert SIMEONI

Jorge ANDRE, Didier CALOGINE, Pedro CUIÑAS, Jean Luc DUPUY,
Paulo FERNANDES, Michel LARINI, Ana Isabel MIRANDA,
Dominique MORVAN, Josep PIÑOL, Olivier SERO-GUILLAUME

December 2006

The views expressed are purely those of the authors and may not, in any circumstances, be regarded as stating an
official position of the European Commission
EUFIRELAB

CONTENT LIST
1 A review on empirical modelling of wildland fire behaviour................................................................................1
1.1 Introduction....................................................................................................................................................1
1.2 Empirical fire behaviour modelling in Australia and Canada.........................................................................2
1.2.1 Australian models.....................................................................................................................................2
1.2.2 Canadian models .....................................................................................................................................3
1.2.3 Input data and functional relationships.....................................................................................................3
1.3 Empirical fire behaviour modelling in Europe................................................................................................4
1.3.1 The different models.................................................................................................................................4
1.3.2 Fire regime models...................................................................................................................................5
1.3.3 Considerations about research priorities..................................................................................................5
1.4 References ....................................................................................................................................................6
2 A review on physical modelling of wildland fire behaviour ...............................................................................10
2.1 Introduction..................................................................................................................................................10
2.2 Semi-empirical models ................................................................................................................................10
2.3 Physical models...........................................................................................................................................10
2.4 Detailed physical and multiphase approach................................................................................................12
2.4.1 The fuel as a multiphase medium ..........................................................................................................12
2.4.2 The fuel as a fractal medium..................................................................................................................12
2.4.3 Changing spatial scale ...........................................................................................................................13
2.4.4 A system of multiphase, reactive and radiative equations .....................................................................13
2.4.5 Using asymptotic analysis......................................................................................................................13
2.4.6 Large Eddy Simulation ...........................................................................................................................14
2.5 Approximate models....................................................................................................................................14
2.6 Future prospects .........................................................................................................................................15
2.6.1 To characterise the fuel..........................................................................................................................15
2.6.2 To validate the fire behaviour.................................................................................................................15
2.6.3 To improve the metrology ......................................................................................................................15
2.6.4 To change the scale of prediction ..........................................................................................................15
2.6.5 To use larger numerical meshes............................................................................................................15
2.6.6 To develop approximate models ............................................................................................................15
2.7 Conclusion...................................................................................................................................................16
2.8 References ..................................................................................................................................................16
3 A review on wildland fire smoke dispersion modelling .....................................................................................20
3.1 Introduction..................................................................................................................................................20
3.2 Emission and dispersion of air pollutants....................................................................................................20
3.2.1 The major air pollutants..........................................................................................................................20
3.2.2 Emission of particulate matter................................................................................................................20
3.2.3 Modification of the mixture .....................................................................................................................21
3.2.4 Transportation of the mixture .................................................................................................................21
3.3 Numerical modelling of smoke production and dispersion..........................................................................22
3.3.1 To estimate the plume rise.....................................................................................................................22
3.3.2 To calculate the transport speed and direction ......................................................................................22
3.3.3 To simulate smoke dispersion................................................................................................................23
3.4 Final comments ...........................................................................................................................................26
3.5 References ..................................................................................................................................................27
3.6 Figures.........................................................................................................................................................31
4 Fire behaviour prediction ..................................................................................................................................33
4.1 Introduction..................................................................................................................................................33
4.2 Fire development phase..............................................................................................................................33
4.3 Fire regime ..................................................................................................................................................34
4.4 Aspects of the fire behaviour predicted.......................................................................................................35
4.4.1 Main fire front .........................................................................................................................................35
4.4.2 Behaviour of a small section of the fire line (local prediction)................................................................35
4.4.3 Behaviour of the whole fire line (global prediction) ................................................................................36
4.4.4 Large crown fires and bushfires .............................................................................................................36
4.4.5 Spotting ..................................................................................................................................................37
4.4.6 Fire whirls ...............................................................................................................................................37
4.5 Applied research products...........................................................................................................................37
4.6 Future prospects .........................................................................................................................................38
4.7 References ..................................................................................................................................................38

D-03-09.doc
EUFIRELAB

SUMMARY
In this document, the authors summarise the two main approaches for modelling the behaviour of wildland fire
classically followed by the scientific community: empirical and physical modelling.
Concerning the empirical modelling, the authors analyse the models developed in Australia since the 1960’s in
the different type of ecosystems concerned by wildland fire: mainly grassland and eucalypt stands, and, more
recently in moorlands and shrublands.
Then analyse also the models developed in Canada related to the Canadian Forest Fire Rating System in
constant evolution since the 1920’s.
The authors pay attention to the required input data of these empirical models and the empirical functional
relationships established between parameters like damping effect of the moisture content, wind and slope effects
on wildland fire dynamic, or flame length and height and wildland fire intensity.
Afterwards, they analyse the empirical approach for modelling the wildland fire behaviour in Europe, in fact in
the Euro-Mediterranean countries.
An original approach, called “fire regime modelling” is also presented. It aims to represent the impact of fires on
a landscape scale over several decades, in terms of finally burnt area.
They add considerations about research priorities in the domain of the empirical approach in relation with the
current state of the art.
Concerning the physical modelling, the authors compare the different physical models, which take into account
one or several processes of energy transfer from the burning zone to the unburned fuel.
The authors describe largely the detailed physical and multiphase approach, indicating the nature of the solved
equations.
Before developing some future prospects, they analyse the interest of approximate models.
The third part of the document is dedicated to a complete review of the smoke dispersion models used
worldwide. The authors classify them in three groups: research models, planning models and screening models.
They underline the interest of fire assessment systems and expert systems in this domain.
In the whole document, the authors present the contribution of the European teams involved in EUFIRELAB.
They are involved in all kinds of modelling and references are included in all sections.
At the end of each part, the authors provide a very large list of bibliographical references concerning the items
developed in the three chapters of the document.

D-03-09.doc
EUFIRELAB

1 A REVIEW ON EMPIRICAL MODELLING OF WILDLAND FIRE BEHAVIOUR

1.1 INTRODUCTION
Finally, yet importantly, various practical,
Reviews of fire behaviour models (BEER 1991a; economical and legal matters strongly restrain the
BURROWS 1994; PERRY 1998; DUPUY 1999; ANDRÉ and number and size of field trials and the range of
VIEGAS 2001; WEBER 2001; PASTOR et al. 2003) usually experimental burning conditions, which are acceptable,
recognize three categories, respectively empirical (or which poses constraints on the performance of the
statistical), semi-empirical (semi-physical or laboratory resulting models.
models), and physical (theoretical or analytical).
E.g., despite its long tradition of experimental
The physical modelling of fire behaviour seeks wildland fires, it was only by the end of the 1990's that
mathematical solutions for the complex mechanisms a vast and well-funded research program focused on
involved in fire propagation. high-intensity crown fires was undertaken in Australia.
A theoretical formulation capable of a direct,
Some of the advantages of empirical fire models are
practical, and truly predictive response to the entire fire
readily apparent, namely:
behaviour variability that can naturally occur is quite
- the absence of artificiality and scale problems
distant (ALEXANDER et al. 1998).
(present in the lab experiments associated to
It will depend on the nature of its inputs (can they be
physically-based models), and
readily acquired in the real world?) as well as on the
- the integration of numerous factors which operate in
availability of powerful computing resources.
the real world, such as wind and moisture profiles
Consequently, fire modelling for operational
(impossible to reproduce in the lab) and fuel
applications is currently restricted to semi-empirical or
heterogeneity.
empirical models.
This is probably why an empirical model developed
Empirical fire behaviour models are based on data
for a given vegetation type usually performs better than
collected in experimental fires or in well-documented
ROTHERMEL's semi-physical model (LINDENMUTH and
wildfires or prescribed fires.
DAVIS 1973; MARSDEN-SMEDLEY and CATCHPOLE 1995;
Although containing an element of uncertainty, the
FERNANDES 1998; VEGA et al. 1998; BURROWS 1999;
former are frequently used to expand the range of
HÉLY et al. 2001).
environmental conditions and fire characteristics.
Therefore, an empirical model can be viewed as a
Empirical models are built by correlating the
fast (and intrinsically incomplete) answer to a specific
observed fire characteristics with easily measured
problem of fire behaviour prediction.
variables, which describe the so-called fire environment
For the time being, the most apparent benefit of a
(fuel, weather and slope).
physical approach to fire modelling is its contribution to
Observational evidences and statistical criteria
understand fire propagation mechanisms, therefore
guide the process, but the functional relationships
helping the experimental design and interpretation of
employed in contemporary models try to be consistent
field trials (BURROWS 1994).
with theoretical knowledge.
Nevertheless, the increasing knowledge on the
Empirical models should not be extrapolated beyond fundamental fire behaviour allows developing simplified
the data range used for their development, but this physical models useful for simulation tools (see
basic rule is seldom respected: it is better to use a sections 2.4 and 2.5).
(probably) poor model than no model at all.
This is not the sole weakness of empirical models,
as environmental factors cannot be controlled in the
field.
As a result, natural or circumstantial correlation
between variables can hide their true effects and
complicate the detection and quantification of
relationships with the dependent variable, e.g. CHENEY
et al. (1993).

D-03-09.doc 1
EUFIRELAB
1.2 EMPIRICAL FIRE BEHAVIOUR MODELLING
IN AUSTRALIA AND CANADA
1.2.1 Australian models
Fire behaviour research in Australia is pragmatic:
the objective is to derive relationships which are
consistent with management necessities (fire danger
rating and accurate prediction of prescribed fire
behaviour, essentially) and are immediately usable,
based in easily acquirable variables.
The existing diversity of climates and vegetation
types has led to a variety of regional models (BURROWS
1994), which classically have been provided to end
users in the form of circular meters (e.g. MCARTHUR
1967) or sequenced tables and graphs (e.g.
SNEEUWJAGT and PEET 1985).
The fire behaviour relationships underlying the first
generation of fire behaviour guides were the product of
experimental fires lit by point ignition in specific fuel
types and mild fire weather, with additional data
collected in wildfires in order to make the predictions
more respondent to severe conditions (CHENEY 1981;
BURROWS 1994).
Fire spread in grassland (MCARTHUR 1966) and dry
eucalypt forest (MCARTHUR 1962, 1967) in
southeastern Australia was predicted from dead fuel
moisture content and wind speed, and then adjusted for
slope.
Forest situations considered also the fine fuel load
and fuel availability (MCARTHUR 1962) or a drought
factor (MCARTHUR 1967).
The importance of other fuel characteristics, namely
the fuel bed compactness and particle size has been
recognised, but their effects were disregarded,
probably due to difficulties in obtaining exact
measurements and the high field variability or simply
because their effects were not quantifiable (CHENEY
1981; TOLHURST and CHENEY 1999).
The Forest Fire Behaviour Tables (FFBT)
(SNEEUWJAGT and PEET 1985) were developed in the
south-west of Australia by similar methods.
The Red Book, as its is often called, applies to 11
fuel complexes one of two models, respectively for dry
eucalypt forest of Eucalyptus marginata with sparse
and low understorey, and for humid eucalypt forest of
Eucalyptus diversicolor with dense and tall understorey.
A basic spread rate is calculated from dead fuel
moisture content and wind speed for a standard fuel
type, and subsequently is adjusted for available fuel
and terrain slope.
The above-mentioned predictive schemes are
currently available as equations, which were fit to the
original tables, graphs and meters by NOBLE et al.
(1980), GOULD (1993), BECK (1995) and GRIFFITHS
(1999) to facilitate calculation and comparison between
systems.
D-03-09.doc
Given their solid data background, these models
perform well when applied to the situations for which
they were developed (CHENEY 1981), but the
agreement between observations and predictions is
usually deficient outside the original range
(UNDERWOOD et al. 1985; TOLHURST and CHATTO 1998;
BURROWS et al. 2000).
The models for eucalypt forest tend to
underestimate the rate of spread of fires initiated by a
line, especially when wind is stronger and shrubs are
taller (GOULD et al. 2001).
According to BURROWS (1994), the explanation lies
not just in the original range of data but also in the
methodology used to establish the relationships: some
were inferred rather than derived directly, and there's a
lack of descriptive and statistic information concerning
the process of model development.
Fire management requirements are increasingly
demanding and, consequently, more accurate and
exact fire behaviour predictions are required.
A second generation of Australian fire models
appeared in the 1990's, carefully examining the existing
assumptions, enlarging the spatial scale and the
meteorological range of experimental conditions, and
based on sophisticated statistical analysis.
Some previously ignored influences were
experimentally addressed, such as the effect of the
ignition line length (CHENEY and GOULD 1995).
CHENEY et al. (1992) and BUCKLEY (1993) presented
equations for eucalypt stands with an important shrub
component.
CHENEY et al. (1993, 1998) qualitatively
individualised two pasture types and introduced
considerable modifications in the wind speed function,
which has resulted in a new meter for grassland
(CSIRO 1997).
The model of BURROWS (1999) for SW Australia
eucalypt forest is much simpler than the cumbersome
Red Book.
In the same forest environment, the VESTA project
has expanded the weather conditions of
experimentation to extreme fire danger levels, as well
as examining in detail the polemic effects of fuel load,
age and structure on spread rate (GOULD et al. 2001).
Fire behaviour models are now available for other
Australian environments, namely discontinuous
vegetation types in arid (BURROWS et al. 1991) and
semi-arid (MCCAW 1991, 1998) regions, buttongrass
moorland (MARDEN-SMEDLEY and CATCHPOLE 1995)
and shrublands (CATCHPOLE et al. 1998).
New approaches include some physics in the
second-generation models. They describe heat
transfers (JOHNSTON et al. 2006) or buoyancy (DOLD et
al. 2006) in order to represent the fire behaviour in
various conditions (fuel heterogeneity or ignition line
length).
2
EUFIRELAB
1.2.2 Canadian models
The Canadian Forest Fire Behaviour Prediction
System (FBP) is part of the Canadian Forest Fire
Danger Rating System, whose concept has
continuously evolved during the 1920's – 1990's period.
The system estimates fire behaviour as a function of
indexes belonging to the FWI (Fire Weather Index) sub-
system.
So, and contrarily to the Australian approach, there
are no direct functional relationships between the
descriptors of fire behaviour and the fire environment.
The FWI sub-system has temperature, relative
humidity, precipitation and wind speed as inputs,
variables, which allow the calculation of six
components (VAN WAGNER 1987; STOCKS et al. 1989):
- Fuel moisture content indexes: Fine Fuel Moisture
Code (FFMC), Duff Moisture Code (DMC) and
Drought Code (DC);
- Fire behaviour indexes: Initial Spread Index (ISI),
Buildup Index (BUI) and Fire Weather Index (FWI),
respectively representing rate of fire spread,
available fuel, and fire line intensity.
The FBP sub-system includes 89 equations, derived
statistically (but with some theoretical background) from
the analysis of an extensive quantitative database of
experimental fires and wildfires (Forestry Canada Fire
Danger Group 1992).
Data from wildfires (e.g. Stocks 1988) is more
general and less reliable, but is important as it
represents extreme events.
Fire behaviour predictions (which include crown fire
propagation) are provided for 16 specific fuel types but,
differently to other empirical models, structural variation
within a vegetation type is not accounted for.
The ISI is the basic input for rate of spread
estimation on each fuel type, followed by corrections for
the BUI and terrain slope.
The FWI sub-system is increasingly being used in
fire danger rating in several countries.
Because of its structure, the FBP system allows
adaptation to fuel types outside Canada, e.g. FOGARTY
et al. (1998).
The FBP database was analysed by CRUZ et al.
(2002) in order to derive a model for the rate of spread
of crown fires in conifer forest.
The independent variables in the equation are wind
speed, canopy bulk density and dead fuel moisture
content.
D-03-09.doc
1.2.3 Input data and functional relationships
A particular type of models tries to describe the
conditions under which a ignition succeeds, i.e. the
environmental thresholds for sustained fire
propagation.
This is because the moisture of extinction
(ROTHERMEL 1972) in a given fuel type is not constant,
and is affected by wind speed and the characteristics of
the fuel complex.
The most common modelling approach is logistic
regression, whose outcome is a probability of fire
sustainability (LAWSON et al. 1994; MARSDEN-SMEDLEY
et al. 2001).
Crown fire initiation models also rely on logistic
models, with wind speed, fuel strata gap, surface fuel
consumption class and dead fuel moisture content as
independent variables (CRUZ et al. 2002) or,
alternatively (CRUZ et al. 2003), with wind speed, crown
base height and combinations of indexes from the FWI
as predictors.
Models for fire spread typically combine in a
multiplicative way functions for the individual effects of
wind speed, fuel moisture and a fuel complex
descriptor, with a subsequent correction for slope.
Wind speed measured or estimated at heights of
10m in open terrain is required for fire spread prediction
by the Canadian system and a few other models
(MCARTHUR 1966, 1967; CHENEY et al. 1998).
The remaining empirical models use surface wind
speed, at the height of 1-2 m.
Wind speeds from weather forecasts (6 or 10 m
height) can be converted to surface wind speed
according to a logarithmic profile (ALBINI and BAUGHMAN
1979) or with specific simple equations (SNEEUWJAGT
and PEET 1985; CHENEY et al. 1992; GOULD 1993;
MARSDEN-SMEDLEY et al. 1999).
Most wind functions describe the wind influence on
rate of spread by power or exponential curves.
The difference is unimportant (BEER 1991b), except
at high wind velocities, where the option for an
exponential equation tends to produce unrealistic high
predictions, e.g. BURROWS (1999).
In the FBP system the wind component of the ISI is
exponential, but rate of fire spread has an upper limit,
usually based in wildfire data.
Few empirical functions are available to describe the
effect of slope terrain in fire propagation (MCARTHUR
1962; VAN WAGNER 1977; CHENEY et al. 1992), possibly
because experiments in flat terrain are prevalent.
All are exponential and provide similar results.
The available information for shrubland suggests a
weaker effect of slope on rate of spread (GREEN 1981;
CATCHPOLE et al. 1999).
The damping effect of dead fuel moisture on fire
spread takes the form of an exponential function in
most models (e.g. CHENEY et al. 1993; MARSDEN-
SMEDLEY and CATCHPOLE 1995; MCCAW 1998;
BURROWS 1999) which is well supported by theory and
combustion experiments in controlled environments,
but some Australian equations assume a linear effect.
3
EUFIRELAB
It is difficult to quantify, under field conditions, the
effect of fuel characteristics on fire spread, and nearly
impossible to distinguish between the influences of
different descriptors without manipulation of the fuel
complex.
The dominant role of wind, slope and moisture, and
natural heterogeneity and correlation between fuel
properties are the main reasons for this problem.
Inputs concerning the physical properties of the fuel
complex generally resort to variables with an apparent
important variation within a fuel type and easily
assessed at the management level.
Rate of spread increases linearly with available fuel
quantity in the FBP system (Forestry Canada Fire
Danger Group 1992) as well as in the eastern Australia
models for eucalypt forest (MCARTHUR 1962, 1967;
CHENEY 1978).
Other models use:
- fuel age (MARSDEN-SMEDLEY and CATCHPOLE 1995),
- vegetation height (CHENEY et al. 1992; CATCHPOLE
et al. 1998),
- a fuel factor calculated from vegetation cover and
patchniness (BURROWS et al. 1991), or
- a qualitative indicator of compactness (CHENEY et al.
1993).
Models for flame length or height and fire intensity
are equally important, because these variables are
directly related with fire suppression difficulty and fire
effects.
BYRAM (1959) was the first to estimate flame length
from a power function in fireline intensity, which by
definition is the product of rate of spread, fuel available
for flaming combustion and fuel heat content.
Various equations fit the model for different fuel
types.
Inverse relations were also developed to estimate
the fireline intensity from flame lengths (ALEXANDER
1982).
In Australia, flame height is generally preferred to
flame length, and is predicted from fire line intensity
(MARSDEN-SMEDLEY and CATCHPOLE 1995) or from the
combination of rate of spread and fuel load in various
linear and non-linear forms (NOBLE et al. 1980; CHENEY
et al. 1992; GOULD 1993; BURROWS 1999).
D-03-09.doc
1.3 EMPIRICAL FIRE BEHAVIOUR MODELLING
IN EUROPE
1.3.1 The different models
The fist attempt at an empirical fire behaviour model
in Europe comes from TRABAUD (1979), who presents
an equation for the spread rate of fires in Quercus
coccifera garigue.
However, few models have been developed up to
now, due to the rarity of field experiments, the
widespread use of ROTHERMEL's model in management
applications, and the focus by several research teams
on physically based models.
The statistical analysis of fire behaviour in Europe
comprehends two types of equations based on field
data, respectively simple descriptions of fire
propagation (e.g. SZCZYGIEL 1988; PEREZ and VALETTE
1995; CARREGA and NAPOLI 1998), and models whose
functional relationships are supported by physical
reasoning whenever possible.
Within-fire analysis of the spatial and temporal
variation of fire behaviour in response to wind and fuel
variation has been examined by FERNANDES et al.
(2000) in shrubland and FERNANDES et al. (in press) in
a Pinus pinaster stand.
This type of studies can give insights on the
development of empirical models, provide detailed data
to test existing models, and is useful to analyse specific
phenomena, e.g. the transition from surface to crown
fire.
The main motivation for empirical fire behaviour
modelling in Europe is, as in other regions of the world,
to acquire a straightforward capacity of predicting fire
behaviour, with input variables easily measured by fire
managers, and using equations which reflect real-world
conditions and produce reliable estimates within the
environmental range of experimentation.
The overwhelming majority of the experiments has
been conducted from Autumn to Spring, which means
the application of the resulting models should be
restricted to predict fire behaviour in the low to
moderate range, namely in the prescribed burning
decision-making process.
Models for shrubland fire behaviour in gorse and
heath communities of NW Spain are presented in VEGA
et al. (1998, 2001), including a slope effect and
covering a relatively wide range in fire behaviour.
A small database from four shrubland types in
Portugal has allowed the development of an interim fire
spread model for no-slope conditions (FERNANDES
2001).
Finally, the most extensive experimentation was
conducted in Pinus pinaster stands in northern
Portugal, and has produced equations to predict the
likelihood of sustained fire propagation and the
characteristics of surface fires in fuel types dominated
by litter, grass-ferns or a low shrub strata (FERNANDES
2002; FERNANDES et al. 2002).
The slope effect included in the head fire rate of
spread model is similar to Australian models.
4
EUFIRELAB
The fuel complex in these models is described by
vegetation height or cover for the purpose of rate of
spread estimation, or by fine fuel load, when the
objective is the prediction of flame length / fireline
intensity.
Current work includes the analysis of a joint
database for shrubland fires from Spain, Portugal,
Australia and New Zealand, which has already allowed
the detection of a live fuel moisture content influence
on rate of spread; such a report is exceedingly scarce
in the literature.
1.3.2 Fire regime models
Fire regime models act on a landscape scale over
several decades.
Their goal is not to reproduce spread rates but
rather the finally burnt area.
They can focus on the interaction between
vegetation and fire regime, often use a cellular
automata or gap structure and are normally spatially
explicit: they are based on a grid structure.
The input data evidently cannot be very detailed and
therefore the use of probabilistic equations is not rare.
Typical output parameters are: annually burnt area, fire
size distribution and stand age distribution.
They do not reproduce individual fires, but long-term
averages of fire size and shape.
Fire regime models can generally be divided into
three modules: vegetation build up, fire propagation
and management.
The simplest form of the vegetation build up module
is to determine for each cell the time elapsed since the
last fire.
More sophisticated models consider biomass
augment, ranging from simple linear increase until
modelling of individual tree growth.
The fire propagation module is normally a
neighbourhood process, i.e. fire spreads from one cell
to its eight adjacent cells.
The propagation from one cell to another is not
generally based on physical laws but probabilistic, and
depends on the amount of accumulated fuel, weather
conditions, slope, etc.
The parameters actually included in the propagation
probability vary very much among the different models;
some models also include spotting.
The management module is used to simulate the
effect of fire suppression and/or extraction of biomass
(mechanical or prescribed fires).
Models of this category often focus on the
interaction between fuel build up and fire.
A few models put more attention to the theoretical
aspects of fire regime itself, using simple vegetation
growing modules on homogeneous terrain without
considering wind and slope (PIÑOL et al. 2005), or
considering them (HARGROVE et al. 2000).
D-03-09.doc
Many concentrate:
- on the effect of fuel heterogeneity on fire patterns,
- pointing out the effect of spatial arrangement of fuel
on crown fires (TURNER and ROMME 1994),
- the effect of connectivity of fuels on surface fire
regimes (MILLER and URBAN 2000),
- how propagation probability scales with stand age
(LI et al. 1997),
- testing a exponential stand age distribution due to
fire (BOYCHUK et al. 1997),
- using vital attributes and fuzzy theory (ROBERTS
1996) or
- on a broad scale, including other disturbances (HE
and MLADENOFF 1999).
Some models include:
- human impacts on fire regime (BAKER 1992),
- the effects of settlement and fire suppression on
landscape structure (FAVIER et al. 2004), on forest-
savanna mosaic (including soil heterogeneity)
(BRADSTOCK et al. 2006) or
- effects of different fire management strategies on
one species (THONICKE et al. 2001)
One recent paper studies the influence of fire on the
global vegetation dynamic equilibrium (BOND et al.
2006)
1.3.3 Considerations about research priorities
The current state of the art of empirical fire
behaviour modelling suggests some strategies and
priorities for the future research activities.
1.3.3.1 To reinforce the connection
To reinforce the connection between empirical and
physical approaches to fire behaviour prediction arises
as the first urgent need.
The partial findings obtained by each method could
orientate the research effort in the complementary
procedure.
We need to have in mind that our aim is to have
models, which can be applied to real situations, and be
useful to the fire-decision makers.
The interaction between both methodological
approaches could accelerate fire propagation
modelling, testing and validation processes.
1.3.3.2 To conduct experimental fires
To conduct experimental fires under a wide range of
conditions.
Up to date most of the field experiments have been
carried out in spring, winter and autumn, in mild
weather conditions.
This limits the applicability of the empirical models to
fire season.
It is necessary to expand the range of the
meteorological variables to reflect typical summer
meteorological scenarios.
At the same time, the range of certain variables like
live fuel moisture needs to be also enlarged.
It is necessary to explore different combinations of
environmental variables (fuel, topography and weather)
to avoid co-linearity between variables, a common
problem in the empirical approach.
5
EUFIRELAB
1.3.3.3 To determine the effect of slope terrain
To determine the effect of slope terrain in fire
propagation and its interaction with wind, to obtain the
maximum fire propagation direction and the rate of
spread for any given direction.
The influence of slope on fire propagation has been
taken account by a few empirical models, but most of
field research has been conducted on flat terrain.
Moreover, the interaction between wind and slope, a
relevant factor in fire propagation, is still poorly
understood.
It is necessary and effort to determine the combined
effect of both variables on fire behaviour.
1.3.3.4 To identify new fuel descriptors
To identify new fuel descriptors with a relevant
influence in fire behaviour and reflecting natural fuels
heterogeneity.
Surprisingly, fuel variables explain a low percentage
of variability in fire rate of spread in the current
empirical models.
Furthermore, sometimes, dummy variables have to
be used to reflect the influence of unknown
characteristics of fuel types.
Fuel inventory and physical fuel particles properties
assessment is an expensive and time-consuming task.
Consequently there is a challenge to find new fuel
variables describing fuel particle array influence and
the natural heterogeneity effects on fire behaviour
variability.
1.3.3.5 To develop empirical models
To develop empirical models for fire behaviour
prediction under different ignition techniques used in
prescribed fire.
The available empirical models have been
conceived to predict fire propagation in a quasi-steady
state with fuels burning freely under natural conditions.
This limits its applicability to prescribed burning
operations where the ignition techniques are a factor of
a relevant influence on fire behaviour.
In fact, ignition pattern tries to avoid that fire
develops a stationary state in equilibrium with
environmental conditions.
Models taking account of the effect of ignition
techniques can be very useful to refine fire prescription
to achieve different management objectives.
1.3.3.6 To implement empirical model
To develop empirical models to predict fire
behaviour in manipulated fuels.
Fuel treatments are commonly used in wild land
areas for a great variety of purposes (fuel breaks,
silviculture activities, wildlife habitat improvement, fuel
accumulation reduction, etc.).
These actions result in a mosaic of fuel complexes
with different characteristics.
These situations are not considered in the existing
empirical models. It is necessary to develop new
models, which take account of these fuel conditions,
and to enable to simulate the potential effect of those
fuel treatments on fire hazard.
D-03-09.doc
1.4 REFERENCES
ALBINI, F.A., and R.G. BAUGHMAN. 1979. Estimating
windspeeds for predicting wildland fire behavior.
USDA For. Serv. Res. Pap. INT-221, Intermt. For.
and Range Exp. Stn., Ogden.
ALEXANDER M.E. 1982. Calculating and interpreting
forest fire intensities. Can. J. of Botany 60: 349-57.
ALEXANDER, M.E., B.J. STOCKS, B.M. WOTTON, and R.A.
LANOVILLE. 1998. An example of multi-faceted
wildland fire research: the International Crown Fire
Modelling Experiment. In Proc. 3rd Int. Conf. on
Forest Fire Research & 14th Conf. on Fire and
Forest Meteorology, VIEGAS, D.X. (Ed.), ADAI. pp.
83-112.
ANDRÉ, J.C., and D.X. VIEGAS. 2001. Modelos de
propagação de fogos florestais: estado-da-arte para
utilizadores. Parte I: introdução e modelos locais.
Silva Lusitana 9(2): 237-265.
BAKER W.L. 1992. Effects of settlement and fire
suppression on landscape structure. Ecology 73:
1879-1887.
BECK, J.A. 1995. Equations for the forest fire behaviour
tables for Western Australia. CALMScience 1(3):
325-348.
BEER, T. 1991a. Bushfire rate-of-spread forecasting:
deterministic and statistical approaches to fire
modelling. Journal of Forecasting 10: 301-317.
BEER, T. 1991b. The interaction of wind and fire.
Bound.-Lay. Meteorol. 54: 287-308.
BOND W.F., WOODWARD F.I., MIDGLEY G.F. 2005. The
global distribution of ecosystems in a world without
fire. New Phytologist 165: 525.
BOYCHUK D., PERERA A., TER-MIKAELIAN M.T., MARTELL
D.L., and LI C. 1997. Modelling the effect of spatial
scale and correlated fire disturbances on forest age
distribution. Ecological Modelling 95:145-164.
BRADSTOCK R.A., BEDWARD M., COHN J.S. 2006. The
modelled effects of differing fire management
strategies on the conifer Callitris verrucosa within
semi-arid mallee vegetation in Australia. Journal of
Applied Ecology 43: 281.
BUCKLEY, A.J. 1993. Fuel reducing regrowth forests
with a wiregrass fuel type: fire behaviour guide and
prescriptions. Dept. of Conserv. and Nat. Res., Fire
Manage. Branch, Res. Rep. No. 40, Victoria.
BURROWS, N.D. 1994. Experimental development of a
fire management model for jarrah (Eucalyptus
marginata Donn ex Sm.) forest. PhD. thesis,
Australian National University, Canberra.
BURROWS, N.D. 1999. Fire behavior in jarrah forest
fuels: 2. Field experiments. CALMScience 3(1): 57-
84.
6
EUFIRELAB
BURROWS, N., B. WARD, and A. ROBINSON. 1991. Fire
behaviour in spinifex fuels on the Gibson Desert
Nature Reserve, Western Australia. J. Arid Environ.
20: 189-204.
BURROWS, N., B. WARD, and A. ROBINSON. 2000.
Behaviour and some impacts of a large wildfire in
the Gnangara maritime pine (Pinus pinaster)
plantation, Western Australia. CALMScience 3(2):
251-260.
BYRAM, G.M. 1959. Combustion of forest fuels. In
Forest Fire: Control and Use, Davis, K.P. (Ed.),
McGraw-Hill, New York. pp. 90-123.
CARREGA, P., and A. NAPOLI. 1998. A quick protocol of
in-the-field measurement for the study of forest and
bush fire spread. In Proc. 3rd Intern. Conf. Forest
Fire Research & 14th Fire and Forest Meteorology
Conf., VIEGAS, D.X. (Ed.), ADAI. pp. 656-582.
CATCHPOLE, W., R. BRADSTOCK, J. CHOATE, L.
FOGARTY, N. GELLIE, G. MCCARTHY, L. MCCAW, L.,
J. MARSDEN-SMEDLEY, and G. PEARCE 1998. Co-
operative development of equations for heathland
fire behaviour. In: Proc. 3rd International Conf. on
Forest Fire Research & 14th Fire and Forest
Meteorology Conf., VIEGAS, D.X. (Ed.), ADAI,
University of Coimbra. pp. 631-645.
CATCHPOLE, W., R. BRADSTOCK, J. CHOATE, L.
FOGARTY, N. GELLIE, G. MCCARTHY, L. MCCAW, L.,
J. MARSDEN-SMEDLEY, and G. PEARCE 1999. Co-
operative development of prediction equations for
fire behaviour in heathlands and shrublands. In
Proc. Australian Bushfire Conf. http: //life. csu. edu.
au / bushfire 99/papers.
CHENEY, N.P. 1978. Guidelines for fire management on
forested watersheds based on Australian
experience. FAO Conservation Guide 4, Rome.
CHENEY, N.P. 1981. Fire behaviour. In Fire and the
Australian Biota, GILL, A.M., R.H. GROVES, e I.R.
NOBLE (Eds.), Australian Academy of Science,
Canberra. pp. 151-175.
CHENEY, N.P., and J.S. GOULD. 1995. Fire growth in
grassland fuels. Int. J. Wildland Fire 5(4): 237-247.
CHENEY, N.P., J.S. GOULD, and I. Knight. 1992. A
prescribed burning guide for young regrowth forests
of silvertop ash. Forestry Commission of New South
Wales, Res. Div., Res. Pap. No. 16, Beecroft.
CHENEY, N.P., J.S. GOULD, and W.R. CATCHPOLE. 1993.
The influence of fuel, weather and fire shape
variables on fire-spread in grasslands. Int. J.
Wildland Fire 3(1): 31-44.
CHENEY, N.P., J.S. GOULD, and W.R. CATCHPOLE.
1998a. Prediction of fire spread in grasslands. Int. J.
Wildland Fire 8(1): 1-13.
D-03-09.doc
CRUZ, M.G., M.E. ALEXANDER, and R.H. WAKIMOTO.
2002. Predicting crown fire behavior to support
forest fire management decision-making. In Proc. IV
Int. Conf. on Forest Fire Research & 2002 Wildland
Fire Safety Summit, VIEGAS, D.X. (Ed.), Luso, 18-23
Nov. Millpress Science Publishers, Rotherdam.
CRUZ, M.G., M.E. ALEXANDER, and R.H. WAKIMOTO.
2003. Assessing the probability of crown fire
initiation based on fire danger indices. For. Chron.
79(5): 976-983.
CSIRO. 1997. The CSIRO Grassland Fire Danger and
Fire Spread Meters.
http://www.ffp.csiro.au/nfm/fbm/grassmet.html.
DOLD J., ZINOVIEV A., WEBER R. 2006. Nonlocal Flow
Effects in Bushfire Spread Rates. In Proc. V Int.
Conf. on Forest Fire Research, VIEGAS, D.X. (Ed.),
Figueira da Foz, 27-30 Nov. Elsevier Publishers.
DUPUY, J.L. 1999. An analysis of semi-empirical and
physical models for fire spread in wildland fuels. In
Wildfire Management, Proc. of the Advanced Study
Course, EFTICHIDIS, G., P. BALABANIS, e A. GHAZI
(Eds.), Algosystems S.A., Athens. pp. 419-438.
FAVIER C., CHAVE J., FABING A., SCHWARTZ D., DUBOIS
M.A. 2004. Modelling forest-savanna mosaic
dynamics in man-influenced environments: effects
of fire, climate and soil heterogeneity. Ecological
Modelling 171: 85-102.
FERNANDES, P.M. 1998. Fire spread modelling in
Portuguese shrubland. In Proc. 3rd Intern. Conf.
Forest Fire Research & 14th Fire and Forest
Meteorology Conf., VIEGAS, D.X. (Ed.), ADAI. pp.
611-628.
FERNANDES, P.M. 2001. Fire spread prediction in shrub
fuels in Portugal. For. Ecol. Manage. 144: 67-74.
FERNANDES, P.M. 2002. Desenvolvimento de relações
preditivas para uso no planeamento de fogo
controlado em povoamentos de Pinus pinaster Ait.
PhD thesis, UTAD. Vila Real.
FERNANDES, P., H. BOTELHO, and C. LOUREIRO. 2002.
Models for the sustained ignition and behavior of
low-to-moderately intense fires in maritime pine
stands. In Proc. IV Int. Conf. on Forest Fire
Research & 2002 Wildland Fire Safety Summit,
VIEGAS, D.X. (Ed.), Luso, 18-23 Nov. Millpress
Science Publishers, Rotherdam.
FERNANDES, P.M., W.R. CATCHPOLE, and F.C. REGO.
2000. Shrubland fire behaviour modelling with
microplot data. Can. J. For. Res. 30: 889-899.
FERNANDES, P.M., C. LOUREIRO and H.S. BOTELHO. In
press. Fire behavior and severity in a maritime pine
stand under differing fuel conditions. Ann. For. Sci.
7
EUFIRELAB

FOGARTY, L.G., H.G. PEARCE, W.R. CATCHPOLE, and
M.E. ALEXANDER. 1998. Adoption vs. adaptation:
lessons from applying the Canadian Forest Fire
Danger Rating System in New Zealand. In Proc. 3rd
Int. Conf. on Forest Fire Research & 14th Conf. on
Fire and Forest Meteorology, VIEGAS, D.X. (Ed.),
ADAI. pp. 1011-1028.
Forestry Canada Fire Danger Group. 1992.
Development and structure of the Canadian Forest
Fire Behavior Prediction System. For. Can. Inf. Rep.
ST-X-3, Ottawa.
GOULD, J.S. 1993. Evaluation of MCARTHUR's control
burning guide in regrowth Eucalyptus sieberi forest.
Aust. For. 57: 86-93.
GOULD, J.S., N.P. CHENEY, and L. MCCAW. 2001.
Project Vesta - research into the effects of fuel
structure and fuel load on behaviour of moderate to
high-intensity fires in dry eucalypt forest: progress
report. In Bushfires 2001 Conf.
GREEN, L. 1981. Burning by prescription in chaparral.
USDA For. Serv. Gen. Tech. Rep. PSW-51, Pacific
Southwest For. and Range Exp. Stn., Berkeley.
GRIFFITHS, D. 1999. Improved formula for the drought
factor in MCARTHUR's forest fire danger meter. Aust.
For. 62(2): 202-206.
HARGROVE W.W., GARDNER R.H., TURNER M.G., ROMME
W.H., DESPAIN D.G. 2000. Simulating fire patterns in
heterogeneous landscapes. Ecological Modelling
135: 243-263.
HE H.S., MLADENOFF D.J. 1999. Spatially explicit and
stochastic simulation of forest-landscape fire
disturbance and succession. Ecology 80: 81–99.
HÉLY, C., M. FLANNIGAN, Y. BERGERON, and D. MCRAE.
2001. Role of vegetation and weather on fire
behaviour in the Canadian mixedwood boreal forest
using two fire behaviour prediction systems. Can. J.
For. Res. 31: 430- 441.
JOHNSTON P., MILNE G., KELSO J. 2006. A heat transfer
simulation model for wildfire spread. In Proc. V Int.
Conf. on Forest Fire Research, VIEGAS, D.X. (Ed.),
Figueira da Foz, 27-30 Nov. Elsevier Publishers.
LAWSON, B.D., O.B. ARMITAGE, and G.N. DALRYMPLE.
1994. Ignition probabilities for simulated people-
caused fires in British Columbia pine and white
spruce-subalpine fir forests. In Proc. 12th Conf. Fire
and Forest Meteorology, SAF Publ. 94-02, Soc. Am.
For., Bethesda. pp. 493-505.
LI C., TER-MIKAELIAN M., PERERA A. 1997. Temporal fire
disturbance patterns on a forest landscape.
Ecological Modelling 99: 137-150
LINDENMUTH, A.W., and J.R. DAVIS. 1973. Predicting fire
spread in Arizona's oak chaparral. USDA For. Serv.
Res. Pap. RM-101, Rocky Mte. For. and Range
Exp. Stn., Fort Collins.
MARSDEN-SMEDLEY, J.B., and W.R. CATCHPOLE. 1995a.
Fire behaviour modelling in Tasmanian buttongrass
moorlands. II. Fire behaviour. Int. J. Wildland Fire
5(4): 215-228.
MARSDEN-SMEDLEY, J.B., T. RUDMAN, A. PYRKE, and
W.R. CATCHPOLE. 1999. Buttongrass moorland fire-
behaviour prediction and management. Tasforests
11: 87-107.
MARSDEN-SMEDLEY, J.B., W.R. CATCHPOLE, and A.
PYRKE. 2001. Fire behaviour modelling in
Tasmanian buttongrass moorlands. IV Sustaining
versus non-sustaining fires. Int. J. Wildland Fire 10:
255-262.
MCARTHUR, A.G. 1962. Control burning in eucalypt
forest. Comm. Aust. For. Bur., Leaflet No 80,
Canberra.
MCARTHUR, A.G. 1966. Weather and grassland fire
behaviour. Aust. Forestry and Timber Bureau,
Leaflet No. 100, Canberra.
MCARTHUR, A.G. 1967. Fire behaviour in eucalypt
forest. Aust. Forestry and Timber Bureau, Leaflet
No. 107, Canberra.
MCCAW, L. 1991b. Fire spread prediction in mallee-
heath shrublands in South-Western Australia. In
Proc. 11th Conf. Fire and Forest Meteorology. pp.
226-233.
MCCAW, L. 1998. Research as a basis for fire
management in mallee-heath shrublands of South-
Western Australia. In Proc. 3rd Inter. Conf. Forest
Fire Research & 14th Conf. Fire and Forest
Meteorology, VIEGAS, D.X. (Ed.), ADAI. pp. 2335-
2348.
MILLER C., URBAN D.L. 2000. Connectivity of forest
fuels and surface fire regimes. Landscape Ecology
15: 145-154
NOBLE, I.R., G.V. BARY, and A.M. GILL. 1980.
MCARTHUR's fire-danger meters expressed as
equations. Aust. J. Ecol. 5: 201-203.
PASTOR, E., L. ZÁRATE, E. PLANAS, J. ARNALDOS. 2003.
Mathematical models and calculation systems for
the study of wildland fire behaviour. Prog. Energy
Comb. Sci. 29: 139-153.
PEREZ, J.Y., and J.C. VALETTE. 1995. L'éclosion et la
propagation initiale du feu dans une lande a molinie
en foret landaise mars a août 1995. INRA, Dept.
Rech. For. Rech. Forest. Mediter., Doc. PIF9525,
Avignon.
PERRY, G.L.W. 1998. Current approaches to modelling
the spread of wildland fire: a review. Prog. Phys.
Geog. 22(2): 222-245.
PIÑOL J, BEVEN K, VIEGAS DX. 2005. Modelling the
effect of fire-exclusion and prescribed fire on wildfire
size in Mediterranean ecosystem. Ecological
Modelling 183: 397-409.

D-03-09.doc 8
EUFIRELAB
ROBERTS D.W. 1996. Landscape vegetation modelling
with vital attributes and fuzzy systems theory.
Ecological Modelling 90: 175-184.
ROTHERMEL, R.C. 1972. A mathematical model for
predicting fire spread in wildland fuels. USDA For.
Serv. Res. Pap. INT-115, Intermt. For. and Range
Exp. Stn., Ogden.
SNEEUWJAGT, R.J., and G.B. PEET. 1985. Forest fire
behaviour tables for Western Australia. 3rd Ed. West.
Austr. Dept. Conserv. and Land Manage., Perth.
STOCKS, B.J. 1988. Forest fire close to home: Terrace
Bay Fire no. 7/86. In Proc. Symp. and Workshop on
Protecting People and Homes from Wildfire in the
Interior West, Fischer, W. e S. Arno (Comps.), USDA
For. Serv. Gen. Tech. Rep. INT-251. pp. 89-93.
STOCKS, B.J., B.D. LAWSON, M.E. ALEXANDER, C.E. VAN
WAGNER, R.S. MCALPINE, T.J. LYNHAM, and D.E.
DUBÉ. 1989. The Canadian Forest Fire Danger
Rating System: an overview. For. Chron. 65(6): 258-
265.
SZCZYGIEL, R. 1988. Forest litter burning in pine stands,
laboratory and field studies. In Proc. Intern.
Workshop on Prescribed Burning, INRA, Avignon.
pp. 137-145.
THONICKE K., VENEVSKY S., SITCH S., CRAMER W. 2001.
The role of fire disturbance for global vegetation
dynamics: coupling fire into a Dynamic Global
Vegetation Model. Global Ecology & Biogeography
10: 661.
TOLHURST, K.G., and K. CHATTO. 1998. Behaviour and
threat of a plume driven bushfire in West-Central
Victoria, Australia. In Proc. 13th Conf. Fire and
Forest Meteorology, IAWF. pp. 321-331.
TOLHURST, K.G., and N.P. CHENEY. 1999. Synopsis of
the knowledge used in prescribed burning in
Victoria. Dept. of Nat. Res. and Environ. Fire
Manage., Melbourne.
D-03-09.doc
TRABAUD, L. 1979. Etude du comportement du feu dans
la garrigue de chêne kermes à partir des
températures et des vitesses de propagation. Ann.
Sci. For. 36: 13-35.
TURNER M.G., ROMME W.H. 1994. Landscape dynamics
in crown fire ecosystems. Landscape Ecology 9: 59-
77.
UNDERWOOD, R.J., R.J. SNEEUWJAGT, and H.G. STYLES.
1985. The contribution of prescribed burning to
forest fire control in Western Australia: case studies.
In Fire Ecology and Management in Western
Australian Ecosystems, Ford, J.R. (Ed.), WAIT
Environmental Studies Group Rep. No. 14. pp. 153-
170.
VAN WAGNER, C.E. 1977. Effect of slope on fire spread
rate. Can. For. Serv. Bi-Mon. Res. Notes 33: 7-8.
VAN WAGNER, C.E. 1987. Development and structure of
the Canadian Forest Fire Weather Index System.
Can. For. Serv., Forestry Tech. Rep. 35, Ottawa.
VEGA, J.A., P. CUIÑAS, M.T. FONTÚRBEL, P. PÉREZ-
GOROSTIAGA, and C. FERNÁNDEZ. 1998. Predicting
fire behaviour in Galician (NW Spain) shrubland fuel
complexes. In Proc. 3rd Int. Conf. Forest Fire
Research & 14th Conf. Fire and Forest Meteorology,
VIEGAS, D.X. (Ed.), ADAI. pp. 713-728.
VEGA, J.A., P. PÉREZ-GOROSTIAGA, P. CUIÑAS, M.T.
FONTÚRBEL, and M.C. FERNÁNDEZ. 2001. Manual de
queimas prescritas para matogueiras de Galicia.
Xunta de Galicia, Consellería de Medio Ambiente,
Centro de Información e Tecnoloxía Ambiental.
WEBER, R.O. 2001. Wildland fire spread models. Pp.
151-169 In Forest Fires Behaviour and Ecological
Effects, E.A. JOHNSON and K. MIYANISHI (Eds.).
Academic Press, San Diego.
9
EUFIRELAB
2 A REVIEW ON PHYSICAL MODELLING OF WILDLAND FIRE BEHAVIOUR
2.1 INTRODUCTION
Forest fire spread models are usually classified into
two types (PITTS 1991):
- Stochastic models consisting to predict the more
probable fire behaviour from average conditions and
accumulating acknowledges obtained from
laboratory and outdoor experimental fires,
- Deterministic models (Semi-empirical and physical)
in which the fire behaviour is deduced from the
resolution of the physical conservation laws (mass,
energy, momentum…) governing the evolution of
the system formed by the flame and its environment,
The main purpose of these models is to predict the
local rate of spread of a fire front, when parameters
characterising the condition of spread (vegetation,
meteorology, terrain) are given.
It is reminded that stochastic models are only based
on the observation of field fires (experimental fires and
wildfires) from which the fire rate of spread (ROS) is
related to relevant parameters in a purely statistical
way (fuel type, fuel loading, fuel moisture, wind, ...).
These empirical relations depend strongly from the
very specific conditions from which the statistical study
was performed.
Without systematic parametric studies, it is very
difficult to extract a general behaviour for the fire.
2.2 SEMI-EMPIRICAL MODELS
Semi-empirical models are based on a global
energy balance (FRANDSEN 1971) and on the
assumption that the energy transferred to the unburned
fuel is proportional to the energy released by the
combustion of the fuel, several terms of the model
being fitted to laboratory fire experimental results
(ROTHERMEL 1972).
The simplicity of this approach has allowed to
develop operational tools such as BEHAVE and
FARSITE.
Other models have been developed in order to
represent the fire-line evolution under slope and wind
conditions (Viegas 2002) or fire blow-ups (Viegas and
Pita 2004).
D-03-09.doc
2.3 PHYSICAL MODELS
Physical models take into account one or several
process of energy transfer from the burning zone to the
unburned fuel.
They describe how a part of the energy released by
the fire preheats the unburned fuel, providing its own
propagation.
The fire front is generally supposed to be an infinite
straight line perpendicular to the direction of spread.
These models are based on a simplified
representation of each of the included phenomena.
Generally, the flame is assimilated as a radiating
panel at a fixed temperature (flame temperature)
transferring its energy toward the vegetation by
radiation.
This leads to writing one equation of energy balance
for the unburned fuel in a frame attached to the flame
(steady-state regime of propagation) and in one space
dimension (the direction of fire spread x).
For this, these models consider a homogeneous
and uniform fuel bed made of one type of particle of
plant material.
The unburned fuel is separated from the burning
zone by a surface where the fuel ignites.
On this surface, named ignition interface, the fuel
particles have been raised to the ignition temperature,
which is a given property of the fuel.
The differences between physical models are
essentially due to the choice of the control volume used
to establish the energy balance equation and due to the
processes of energy transfer they take into account.
Most of these models calculate the energy balance
in one dimension (x), and consider:
- either a control volume over the whole fuel bed
depth (e.g. HOTTEL et al. 1965) or
- a small control volume located at the top of the fuel
bed (e.g. HOTTEL et al. 1965).
ALBINI (1981,1985,1986):
- calculates the energy balance in two space
dimensions (x, z) and
- uses a small control volume around a point of co-
ordinates (x,z) located inside the fuel bed.
In the most sophisticated models of this type, the
energy transfers for preheating the unburned fuel are:
- the radiation from embers (often calculated as
radiation from the ignition interface),
- the radiation from the flame, the radiation from the
unburned particles themselves (ALBINI 1985, 1986),
and
- one or several terms of a convective type rendering
the thermal exchange between the fuel particles and
the gas (PAGNI and PETERSON 1973).
10
EUFIRELAB

Some authors add a term of energy loss to the The balance equation is:
ambient medium, which can be of radiative or - sometimes one-dimensional and time dependent
convective nature (or both). (CEKIRGE 1978),
A recent approach provides the main characteristics - often two-dimensional and time dependent – x, y on
of a 3D fire front (Balbi et al. 2005). a plane or ξ,η on a generalised surface representing
Some empirical relations are combined with mass, a terrain (WEBER 1991a, BALBI and SANTONI 1999).
momentum and energy balances to obtain a very
simple quasi-analytical model (after making some They can also take into account slope and wind
strong assumptions). conditions (SANTONI et al. 1999, SIMEONI et al. 2001,
The model provides: MORANDINI et al. 2003, 2005).
- the fire rate of spread, The energy balance equation may be coupled with a
- the temperature in the fuel layer and model of atmospheric air flow (FERRAGUT et al. 1996,
- the fire front geometry: flame tilt angle, flame height LYMBEROPOULOS et al. 1998).
and fire front width.
It has to be validated at the field scale. Using these models, if the empirical coefficients that
calibrate each term representing each physical process
Hence, physical models often permit a simple in the balance equation, are known, one can calculate
calculus of the fire rate of spread, however only when the fire rate of spread and a temperature field,
some properties are given as input parameters: sometimes with the help of hard numerical methods
embers, flame, hot gases at the vicinity of the ignition (SANTONI 1998, SIMEONI et al. 2002).
interface.
These properties directly depend on the fire itself, Up to this point, none of the reviewed models took
thus they are not parameters, which only characterise into account the basic phenomena, which explain the
the combustible medium and the environmental behaviour of a forest fire (pyrolysis, oxygen transport,
conditions. gaseous phase combustion, embers combustion).
It is a weak point of this modelling approach, which Starting from a description of these basic
will be now called a phenomenological approach. phenomena, the properties of the flame, the embers
Nevertheless, it represents the first generation of and the hot gases must be known.
models that are at the root of physical modelling. In others words, the only input parameters of the
model must be the initial characteristics of the medium;
For better understanding the difference with other the model itself calculates these properties.
approaches, it is worth noting that in the previous
models, the energy balance does not contain any term This family of models, based on a detailed physical
of energy source due to the combustion of the plant and multiphase approach, is presented in the following
material. chapter.
This kind of term is implicitly contained into the
properties of the flame, the embers or the hot gases,
used to calculate the energy transfers.
Other models are, in their principle:
- similar to the models previously described and
- planned to be integrated in operational management
tools.
This kind of simulators necessitates simple and
robust models, which are able to provide information on
the fire spread, within a given margin of error and with
a short calculation time.
They consist in using one equation of energy
balance, generally applied to a surface (terrain
surface), which is of a generic mathematical form and
may include convective, diffusive and/or radiative
processes.
In addition, the balance equation contains a term of
energy source, which often depends on the
temperature.
It describes the consumption of fuel but without any
reference to a chemical reaction (global heat release).

D-03-09.doc 11
EUFIRELAB
2.4 DETAILED PHYSICAL AND MULTIPHASE
APPROACH
The main groups of work based on a detailed
physical and multiphase approach of forest fire
behaviour modelling found in the literature are
described hereafter.
2.4.1 The fuel as a multiphase medium
GRISHIN (1997) reports in a book the main results of
the researches carried out in Russia (and ex-USSR) on
the modelling of forest fire behaviour (see also GRISHIN
et al 2002).
Among these works, a general mathematical model
of forest fire behaviour has been described at a first
time in 1992.
The numerical solutions of this detailed model are
not presented.
However, numerous approximate models based on
this approach have been produced before this date and
are described in the same book (GRISHIN 1997).
LINN (1997) presents a PhD dissertation entitled “a
transport model for prediction of wildfire behaviour”,
which is a first step of a wider project aiming to develop
the abilities of “self-determined” models of fire
propagation through forest fuels.
The practical interest of the physical approach to
study wildfire behaviour was demonstrated in a paper
(HANSON et al 2000) comparing:
- numerical predictions obtained for historical fires
using various physical formulations
- with those obtained with classical empirical (or semi-
empirical) approach.
Finally, LARINI et al. (1998) (and in a more teaching
form LARINI, 1998), PORTERIE et al (2000), MORVAN et al
(2000,2001,2002,2003) provide the bases for the
formulation of a complete model for forest fire
propagation.
They describe how the usual equations of
continuum media mechanics can be transformed
following a rigorous method into equations well-suited
to a multiphase medium, a vegetation sustaining a fire.
We summarise here after the last two approaches,
because only these authors completely describe how
the multiphase equations for forest fire modelling are
obtained.
Then, we briefly mention the main differences with
the works of Grishin and Linn.
When a combustible medium is observed at a
sufficiently small scale, a gaseous phase that
surrounds solid particles of combustible vegetation can
be seen.
These particles can be leaves, small twigs, needles,
grasses, ...
The combustible medium is considered as a
multiphase medium composed of a gaseous phase and
several solid phases.
D-03-09.doc
Each solid phase consists of the same fuel particles,
which have the same shape, the same size, the same
physical-chemical properties, thus the same behaviour
with respect to a fire.
In particular, each family of particle is characterized
by its Surface Area / Volume (SA/V) ratio.
In order to achieve a spatial description of the
medium, the fraction of space volume occupied by
each family or each solid phase, at each point and
each time, must be known.
2.4.2 The fuel as a fractal medium
In the approach of SÉRO-GUILLAUME et al. (2002),
the vegetal phase is considered as a fractal medium,
which can be described mathematically.
Moreover, the different parts of plants are
considered as a porous medium, in order to take into
account the internal structure of vegetation.
Therefore, this model is more complex than the
preceding one, and the difference of viewpoint leads to
substantial change in the mathematical formulation.
If the medium is known, what are the mechanisms
that will govern the propagation of a fire through this
medium?
Observing the medium at a very small scale, which
allows distinguishing one fuel particle from surrounding
gas, the behaviour of this particle is described in the
case of an approaching fire:
- The particle receives energy from this fire by
convection and radiation.
- Its temperature is then raised from the ambient
temperature to the boiling water temperature, at
which it looses its water.
- As soon as the particle is dried, its temperature can
raise again.
- The plant material is decomposed gradually
releasing pyrolysis gases.
- The reactive part of these gases is combined with
the oxygen of the air (gaseous phase chemical
reactions).
- The gas temperature then increases, causing its
expansion and, due to buoyancy forces, the gas
moves.
- These movements play a very important role in the
necessary transport of oxygen and in the energy
transfers.
- At the end of the pyrolysis, the particle is mainly
composed of chars.
- Oxygen may reach the surface of the particle and
react with the remaining chars.
- It’s the combustion of chars, which causes the
regression of the particle surface.
Now the problem is to know how to describe the
evolution of such a medium at the scale of a particle.
The medium is known when at each point of the
gas, the velocity, the temperature, the mass fraction of
chemical species and the pressure are known, and
when at each point of the motionless particle, the
temperature, the mass fraction of chemical species and
the density are known.
12
EUFIRELAB
At this scale, for calculating the state of the medium:
- at each point of the gas, the balance equations of
reactive fluid mechanics are available: mass,
momentum, energy and chemical species, and
- at each point of the particle, a system of balance
equations (mass, chemical species, energy) is
available, which describes the behaviour of a solid
conductive medium that can loose its water and be
decomposed into pyrolysis products.
2.4.3 Changing spatial scale
In addition, at the interface between the solid
particle and the gas, some conditions must be verified
(local interface conditions).
Solving all these equations at this scale can be done
only in very simple geometric configurations, but cannot
be done in real configurations with numerous particles
distributed in a gas.
In this last case, and at this scale, it is impossible to
obtain solutions of the strongly coupled balance
equations at each point of the gas, each point of each
particle, which in addition must verify all local interface
conditions.
Thus, it is necessary to increase this scale of
observation.
From a mathematical point of view, operating this
change of scale is equivalent to average the equations
established at a point of the gas or at a point of the
particle over some volume around these points.
It is a usual operation for modelling multi phase
multi constituent materials and several methods of
averaging are available.
In both approach, using a spatial weighting function
(ANDERSON and JACKSON 1969), balance equations for
the gaseous phase and balance equations for each
solid phase have been integrated over:
- the whole volume occupied by the gas, and
- the whole volume occupied by each solid phase.
After some mathematical transformations, a system
of equations is obtained for each phase (gaseous
phase and solid phases).
These equations establish the relations in time and
space between the weighted average of the different
variables (temperature, velocity, density,...).
2.4.4 A system of multiphase, reactive and
radiative equations
It is worth noting that the averaging operation leads
to the appearance of new terms in the balance
equations, which are due to the interactions between
the different phases (here between the gaseous phase
and each solid phase).
Thus, systems of partial derivative equations are
obtained that are strongly coupled through these terms
of interaction.
For instance, these terms of interaction are due to
the mass transfers related to:
- the drying and pyrolysing process of the particles,
- the drag forces,
- the heat transfers by convection and by radiation.
D-03-09.doc
A system of multiphase, reactive and radiative
equations, rigorously deduced from the instantaneous
point equations (generalised NAVIER-STOKES equations
for the gas and point balance equations for the solid), is
now available.
Before solving these equations, it is necessary to
close them.
The closure operation differs for both approaches.
LARINI’s approach (1998) made some assumptions
and sub-models describing the basic phenomena at the
multiphase scale are provided.
Among these phenomena, there are chemical
reactions and different interactions between phases,
which appear after the transformation of point
equations into multiphase (i.e. averaged) equations.
In the other approach, closure is obtained by
thermodynamics argument.
Let us pinpoint that the used thermodynamic is an
extended one allowing the averaging on very large
representative element volume REV; the concept of
REV is equivalent to the notion of fluid particle.
For this, it is necessary to make some assumptions
and furthermore, to establish sub-models describing
the basic phenomena at the multiphase scale.
These systems of partial derivative equations have
been yet solved in one and two space dimensions
(plane symmetry and axi-symmetry) (MORVAN et al.
1998, PORTERIE et al. 1998, 2000).
These approaches have been used with success to
study various fire propagating through:
- a dead fuel bed (PORTERIE et al. 1998, 2000)
(MORVAN et al. 2000, 2001),
- a living Mediterranean shrub (MORVAN et al. 2002),
- a complex shrub/canopy Mediterranean forest,
- a fuel break (MORVAN et al. 2003) (EFAISTOS and
FIRESTAR research programmes).
Some physical properties such as the ratio between
the respective contribution of the two modes of heat
transfer (radiation and convection) between the flame,
the burned hot gases and the vegetation has been
quantified for the two regimes of propagation of a
surface fire: plume-dominated and wind-driven fires.
Calculating numerical solutions of these equations is
hard and time consuming.
This is the reason why simplified models have been
also deduced from the complete sets of equations
(GIROUD 1997, DUPUY and LARINI 1999).
2.4.5 Using asymptotic analysis
MARGERIT et al. (1997, 2002), proposed a new
approach using in order to derive or relate the detailed
physical approach to 2D propagation models.
They obtained a model very similar to the so-called
“physical model”, see above.
Propagation of fire is modelled by a reaction
diffusion system of partial differential equations.
This model has been coded and numerical
experiments of propagation have been provided at
laboratory and terrain scales.
13
EUFIRELAB
This effort is continuing and recent results show the
clear relation between different approaches of
modelling: SÉRO-GUILLAUME et al. (2002), SÉRO-
GUILLAUME (2003), INFLAME and SPREAD research
programmes.
The multiphase approach has also been used to
improve phenomenological models (SIMEONI et al.
2001) and to develop simplified flow sub-models
(SIMEONI et al. 2003).
2.4.6 Large Eddy Simulation
A recent approach is worthwhile to mention. It
involves “Large Eddy Simulation” LES modelling to
describe the turbulence in the fire plume.
It has been developed stating that the evolution of
large eddy structures characteristic of most fire plumes
is lost with the classical Reynolds-Average Navier-
Stokes RANS methods.
The application of LES techniques to fire is aimed at
extracting greater temporal and spatial fidelity from
simulations of fire performed on the more finely
meshed grids allowed by ever-faster computers.
The assumptions made to model the turbulence
must ultimately be justified by comparison to
experiments.
A first model is the extension to wildland fires of the
“Fire Dynamics Simulator” FDS (MCGRATTAN 2004)
developed by the NIST for fire-driven fluid flow in
buildings and industry.
This new simulation tool is called WFDS (MELL et al.
2006a and b).
The LES approach was:
- applied to crown fire initiation (TACHAJAPONG et al.
2006) and
- validated by experiments at the laboratory scale.
D-03-09.doc
2.5 APPROXIMATE MODELS
Thus, two kinds of approximate models are
available: those of the phenomenological approach
presented at the beginning of this review, and those,
which are directly deduced from the complete physical
and multiphase approach, described above.
Linn (1997) aims at simulating large-scale forest
fires and introduces a sophisticated turbulence model,
including the dissipation of turbulent energy due to the
solid particles.
However, because of the large scale of space, he
had to make strong assumptions on some of the basic
mechanisms (especially the combustion process in the
gaseous phase is not really described).
The approach remains very similar to the previous
one (LARINI et al 1998).
In the different models proposed by GRISHIN (1997),
there is only one solid phase made of different
components (dry material, water, chars, ashes).
In addition, for most of these models, the solid
phase is assumed to be in thermal equilibrium with the
gas (one-temperature models).
Hence, this approach differs from the previous ones.
Physical and semi-empirical approaches are
complementary ones; they can be mixed to describe
wildfire behaviour at large scale (see CRUZ et al. 2006 a
and b for crown fires).
Recent studies have shown the interest to compute
the local atmospheric flow using a physical approach
coupled with a heat released curve representing the
fire.
In this case the ROS of the curve fire is evaluated
from a modified ROTHERMEL model taking into account
the local conditions of propagation (state of the
vegetation, wind velocity, slope…) (CLARK et al. 1996,
HANSON et al. 2000).
14
EUFIRELAB
2.6 FUTURE PROSPECTS
In terms of future prospects, we want to underline
that every modelling approach must be accompanied
by experiments.
The reverse is true, except if one chooses to follow
a statistical or purely empirical approach of the
phenomena.
2.6.1 To characterise the fuel
First, it is necessary to characterise the fuel:
- on one hand the particles, which compose the
medium (morphological and physical-chemical
properties), and
- on the other hand, the combustible medium, in other
words the spatial arrangement of the particles.
One attempt has been done (CALOGINE et al. 2001,
2002) to establish models describing the vegetation in
details from parameters that should be easy to obtain
on the field, this work should be pursued and improved,
see report of WP 02.
2.6.2 To validate the fire behaviour
Second, of course, experiments must permit the
validation of fire behaviour models, by carrying out both
laboratory and field fires.
Conversely, this kind of experiment can show
phenomena, which cannot be easily explained, and
thus make questions to the modellers arise.
This kind of experiment is yet widespread at the
international level (studies of wind effect, slope effect,
fuel type effect... on fire behaviour).
But, faced with the increasing requirement of varied
validation data, these experiments should be
encouraged.
Finally, in particular in the frame of the complete
physical and multiphase approach, the closure of forest
fire behaviour models requires physical sub-models,
which can be directly established only with the help of
specific laboratory experiments.
In this case, the experiment must sufficiently isolate
the studied physical mechanism.
Up to now, this kind of experimental work has been
scarce in the forest fire-modelling domain, and it is
necessary to develop them.
This lack has been partially addressed in the frame
of the current project (see the deliverables of Unit 7
“Metrology”).
D-03-09.doc
2.6.3 To improve the metrology
All these experiments must be accompanied by an
improvement of the metrology, in both laboratory and
field conditions.
For instance, gas temperature measurements using
thermocouple sensors have shown limits, and other
physical variables are still difficult to be measured:
- field of gas velocity,
- solid phase temperature,
- gas and particles radiation
It seems that optical methods should be a relevant
solution to numerous arising problems (see also unit 7).
2.6.4 To change the scale of prediction
Today, the computer resources, which are available
for solving the partial derivative equations derived from
the detailed physical and multiphase approach, permit
the correct calculation of the physical phenomena only
at the local fire scale (~100-200 m) and during a
relatively short time (~10 minutes), and in addition, with
still high duration of calculation.
The objective is to solve these equations in larger
spatial domains, in three space dimensions, and more
generally, to change the scale of prediction.
For this, the most obvious way is to use “ super-
computers ” (e.g. massively parallel computers) with
increasingly efficient numerical methods (algorithms).
2.6.5 To use larger numerical meshes
Another idea is to use larger numerical meshes
without transgressing too much the physics of the
model.
This has been initiated by LINN who used PDF
(Probability Density Function) for describing the source
terms of equations.
This kind of work must continue, but new research
ways, which would be more specific of the forest fire
problem, should be investigated.
2.6.6 To develop approximate models
In order to change of scale of prediction, developing
approximate (simplified) models can be also envisaged;
some examples are described in section 2.4.
We remind that these models can be either
phenomenological models or models directly deduced
from the complete physical and multiphase approach.
Their interest is that they have a physical base and
that they are easier to solve and less time consuming
than complete models.
Of course, in order to achieve such models, it is
necessary to often make strong hypothesis.
Thus, in order to evaluate the relevance of this
hypothesis in a given studied situation, the detailed
approach may be used.
15
EUFIRELAB
2.7 CONCLUSION
In the face of the thematic diversity and the extent of
the work to carry out, it is essential to group together
and co-ordinate the research means, and in particular,
to not forget that modelling and experimental works are
closely related.
The work to carry out implies long-term researches.
For basic reasons that cannot be developed here,
one can expect that researches conducted in the field
of forest fire behaviour modelling bring up very relevant
scientific questions and leads to producing results of a
more general significance.
However, it will be necessary to reply to some of the
questions of the users of research results as soon as
possible.
For this, it seems that users and researchers should
define together:
- which applied problems must be investigated in
priority and
- which researches are expected to provide the best
solutions.
Indeed, these studies will:
- directly answer to the questions of users, and
- consist in an essential step to progress toward a
complete solution of the problem.
D-03-09.doc
2.8 REFERENCES
ALBINI, F.A. 1981. A model for the wind-blown flame
from a line fire . Combustion & Flame 43:155-174.
ALBINI, F.A. 1985. A model for fire spread in wildland
fuels by radiation. Combustion Science and
Technology 42:229-258.
ALBINI, F.A. 1986. Wildland fire spread by radiation - a
model including fuel cooling by natural convection.
Combustion Science and Technology 45:101-113.
ANDERSON, T.B. and JACKSON R. 1967. A fluid
mechanical description of fluidised bed. Industrial
Engineering Chemistry Fundamentals 6:527-539.
ANDERSON, H.E. 1969. Heat transfer and fire spread.
USDA Forest Service, Research Paper (INT-69),
Ogden, Utah, 20 pages.
BALBI, J.H., P.A. SANTONI,. DUPUY J.L 1999. Dynamic
modelling of fire spread across a fuel bed.
International Journal of Wildland Fire 9 275-284.
BALBI J.H., ROSSI J.L., MARCELLI T., SANTONI P.A. 2005.
Propagation de feux de végétation : un modèle 3d
temps réel. In proc. of Journées Internationales de
Thermique, Tanger.
CALOGINE D, SÉRO-GUILLAUME O, 1998, Characteri-
sation of Vegetation in Forest Fire Models, 3rd
International Conference on Forest Fire Research,
Coimbra 16-20 November.
CALOGINE D, SERO-GUILLAUME O., 2001, Sur un modèle
fractal de végétation, application à la modélisation
de la lutte contre les feux de forêts, Récents progrès
en génie des procédés, Vol. 15, pp. 11-18.
CALOGINE D, SERO-GUILLAUME O., 2002a, L’équation de
NAVIER-STOKES avec mémoire et le transfert de
quantité de mouvement dans un milieu poreux,
Comptes Rendus à l’Académie des Sciences
Mécanique, Vol. 330, pp. 383-389.
CALOGINE D, SÉRO-GUILLAUME O., 2002b, Air flow
model in a tree crown”, IV International Conference
on Forest Fires Research, Coimbra.
CATCHPOLE, E.A. and N.J. DE MESTRE. 1986. Physical
models for a spreading line fire. Australian For.
49:102-111.
CEKIRGE, H.M. 1978. Propagation of fire fronts in forest.
Comp. and Maths. with Appls. 4:325-332.
CLARK T.L., M.A. JENKINS, J. COEN, D. PACKHAM 1996,
A Coupled Atmospheric-Fire Model: Convective
Feedback on Fire Line Dynamics, J. Appl. Meteor,
Vol.35, pp.875-901.
CRUZ, M.G., BUTLER, B.W., ALEXANDER, M.E.,
FORTHOFER, J.M., WAKIMOTO, R.H. 2006a.
Predicting the ignition of crown fuels above a
spreading surface. fire I: Model idealization. Int. J.
Wildland Fire 15, 47-60.
16
EUFIRELAB
CRUZ, M.G., BUTLER, B.W., ALEXANDER, M.E. 2006b.
Predicting the ignition of crown fuels above a
spreading surface fire. Part II: Model behavior and
evaluation. Int. J. Wildland Fire 15, 61-72.
DE MESTRE, N.J., E.A. CATCHPOLE, D.H. ANDERSON,
and R.C. ROTHERMEL. 1989. Uniform propagation of
a planar fire front without wind. Combustion Science
and Technology 65:231-244.
DUPUY, J.L. 1997. Towards a better understanding and
prediction of forest fires propagation : LARINI, M., F. GIROUD, B. PORTERIE, and J.C. LORAUD.
experimentation, testing and model development (In
French). Thesis (Doctorate of University). University
Claude-Bernard, Lyon-I, 272p.
DUPUY, J.L. and M LARINI 1999. Fire spread through a
porous forest fuel bed : a radiative and convective
model including fire-induced flow effects. Accepted
in the International Journal of Wildland Fire.
EMMONS, H.W. 1964. Fire in the forest. Fire Research
Abstracts and Reviews 5:163-178.
FERRAGUT, L.F., ASENSIO M.I., MONTENEGRO R., PLAZA
A., WINTER G., SERON F.J. 1996. A model for fire
simulation in landscapes. Computational Fluid
Dynamics, pp 111-116.
FERRAGUT, L.F., TIMÒN S.M., ASENSIO M.I. 2006.
Radiation modelling in fire spread. In Proc. V Int.
Conf. on Forest Fire Research, VIEGAS, D.X. (Ed.),
Figueira da Foz, 27-30 Nov. Elsevier Publishers.
FRANDSEN, W.H. 1971. Fire spread through porous
fuels from the conservation of energy. Combustion
and Flame 16:9-16.
FUJII, N., J. HASEGAWA, L. PHALLOP, and Y. SAKAWA.
1980. A non-stationary model of fire spreading.
Appl. Math. Modelling 41:76-180.
GIROUD, F. 1997. Contribution to fire propagation
modelling : multiphase approach of forest fires,
development of a propellant fire in a semi-confined
medium (In French). Thesis (Doctorate of
University). University of Provence, IUSTI, Aix-
Marseille-I, 198 p.
GRISHIN A.M., O.V. SHIPULINA 2002. Mathematical
model for spread of crown fires in homogeneous
forests and along openings. Combustion, Explosion
& Shock Waves, Vol. 38(6): 622-632p.
GRISHIN, A.M. 1997. Mathematical modelling of forest
fires and new methods of fighting them. Publishing
House of the Tomsk State University, Russia.
390 pages.
HANSON H.P., M.M. Bradley, J.E. BOSSERT, R.R. LINN,
L.W. YOUNKER 2000, The potential and promise of
physics-based wildfire simulation, Environmental
Science & Policy Vol. 3: 161-172p.
HOTTEL, H.C., G.C. WILLIAMS, and F.R. STEWARD. 1965.
The modelling of fire spread through a fuel bed.
Tenth Symposium (International) on Combustion,
the Combustion Institute, Pittsburgh, 1965, pages
997-1007.
D-03-09.doc
HOTTEL, H.C., G.C. WILLIAMS, and G.K. KWENTUS.
1971. Fuel preheating in free-burning fires.
Thirteenth Symposium (International) on
Combustion, U.S.D.A. Forest Service, Washington,
D.C. The Combustion Institute, Pittsburgh, 1971,
pages 963-970.
LARINI, M. 1998. The complete physical models. Short
Course on Forest fire Behaviour Modelling. Luso-
Coimbra, Portugal, 21-22 November 1998.
1998. A multiphase formulation for fire propagation
in heterogeneous combustible media. International
Journal of Heat and Mass Transfer 41(6-7):881-897.
LINN, R.R. 1997. A transport model for prediction of
wildfire behaviour. PhD dissertation, New Mexico
State University, Dept. of Mechanical Engineering,
Las Cruces, New Mexico; also published as Los
Alamos National Laboratory Report, number LA-
13334-T. 195p.
LINN, R.R. and F. HARLOW 1998. Use of transport
models for wildfire behaviour simulations.
Proceedings of the 3rd International Conference on
Forest Fire Research, Coïmbra, Portugal. Vol. 1:
363-372p.
LYMBEROPOULOS, N., TRYFONOPOULOS, and F.C.
LOCKWOOD 1998. The study of small and mesoscale
wind field. Proceedings of the 3rd International
Conference on Forest Fire Research, Luso,
Coïmbra, Portugal. Vol 1 pp405-418.
MARGERIT J., O. SERO-GUILLAUME 2002. Modelling
forest fires. Part 2 :Reduction to two-dimensional
models and simulation of propagation. Int. J. of Heat
& Mass Transfer Vol. 45: 1723-1737p.
MCGRATTAN, K.B. 2004. Fire Dynamics Simulator
(Version 4), Technical Reference Guide, NISTIR
Special Publication 1018, McGrattan, K., editor
http://fire.nist.gov/bfrlpubs/
MELL, W.E., JENKINS, M.A., GOULD J., CHENEY, P.,
2006a. A Physics-Based Approach to Modeling
Grassland Fires. International Journal of Wildland
Fire, in press.
MELL, W.E., MANZELLO S.L., MARANGHIDES A. 2006b.
Numerical Modeling of Fire Spread through Trees
and Shrubs. In Proc. V Int. Conf. on Forest Fire
Research, VIEGAS, D.X. (Ed.), Figueira da Foz, 27-
30 Nov. Elsevier Publishers.
MORANDINI F., P.A. SANTONI, J.H. BALBI 2001. The
contribution of radiant heat transfer to laboratory-
scale fire spread under the influences of wind and
slope. Fire Safety Journal. Vol. 36: 519-543p.
MORANDINI F., SIMEONI A., SANTONI P.A, BALBI J.H.
2005. A model for the spread of fire across a fuel-
bed incorporating the effects of wind and slope.
Combustion Science and Technology 177, 1381-
1418.
17
EUFIRELAB
MORVAN D., J.L. DUPUY 2001. Modelling of fire spread
through a forest fuel bed using a multiphase
formulation. Combust. & Flame, Vol.127: 1981-
1994p.
MORVAN D., J.L. DUPUY 2002. Numerical simulation of
the propagation of a surface fire through a
Mediterranean shrub. 7th International Symposium
on Fire Safety Science, Boston-Worcester, Vol. 7:
557-568p.
MORVAN D., J.L. DUPUY 2003. Wildfire spread
simulation through a Mediterranean shrub in the
urban/forest interface. Proc. Multiphase and
Complex Flow Simulation for Industry, Cargèse 13p.
MORVAN D., V. TAULEIGNE 2002. Wind effects on
wildfire propagation through a Mediterranean shrub.
Proc. 4th International Conference on Forest Fire
Research, Coimbra 14p.
MORVAN D., V. TAULEIGNE, J.L. DUPUY 2002. Flame
geometry and surface to crown fire transition during
the propagation of a line fire through a
Mediterranean shrub. Proc. 4th International
Conference on Forest Fire Research, Coimbra 11p.
MORVAN, D, J.L. DUPUY, M. LARINI 2000. Multiphase
formulation applied to the modeling of fire spread
through a forest fuel bed. Proc. 28th Symposium
(International) on Combustion, The Combustion
Institute, Vol. 28: 2803-2809p.
MORVAN, D., B. PORTERIE, M. LARINI and J-C LORAUD
1998. Numerical simulation of the behaviour of
turbulent diffusion radiating in a cross wind during a
surface fire. Proceedings of the 3rd International
Conference on Forest Fire Research, Coïmbra,
Portugal. Vol. 1: 389-404p.
MORVAN, D., M. LARINI 2001. Modelling of one
dimensional fire spread in pine needles with
opposing air flow. Combust. Sci. & Tech., Vol.164:
37-64p.
PAGNI, P.J., and T.P. PETERSON. 1973. Flame spread
through porous fuels. 14th Symposium
(International) on Combustion, U.S.D.A. Forest
Service, Washington, D.C. The Combustion
Institute, Pittsburgh, pages1099-1107.
PITTS, W.M. 1991. Wind effects on fires, Prog. Energy
Combust. Sci., 17: 83-134.
PORTERIE B, D. MORVAN, J-C LORAUD and M. LARINI
2000. Fires spread through fuel beds: modelling of
wind-aided fires and induced hydrodynamics.
Physics of Fluid, Vol.12(7): 1762-1782p.
PORTERIE, B., D. MORVAN, M. LARINI and J-C LORAUD
1998. A multiphase model for predicting line fire
propagation. Proceedings of the 3rd International
Conference on Forest Fire Research, Coïmbra,
Portugal. Vol 1: 343-360p.
RAMEZANI S, CALOGINE D, SÉRO-GUILLAUME O. January
2004, Numerical Model of Combustion of
Vegetation, Line and Point Ignition. Deliverable D.
2.12,. Spread Report.
D-03-09.doc
ROTHERMEL, R.C. 1972. A mathematical model for
predicting fire spread in wildland fuels. USDA Forest
Service, Research Paper (INT-115), Ogden, Utah,
40 pages.
SANTONI P.A. 1998. Elaboration of an Evolving
Calculation Domain for the Resolution of a Fire
Spread Model. Num. Heat Transfer 33, 279-298.
SANTONI, P.A., BALBI J.H. 1998. Numerical simulation of
a fire spread model. Proceedings of the 3rd
International Conference on Forest Fire Research,
Luso, Coïmbra, Portugal. Vol 1 pp295-310.
SANTONI, P.A., BALBI J.H., DUPUY J.L. 1999. Dynamic
modelling of upslope fire growth. International
Journal of Wildland Fire 9, 285-292.
SIMEONI A., SANTONI P.A., LARINI M. BALBI J.H. 2001.
Proposal for theoretical contribution for improvement
of semi-physical forest fire spread models thanks to
a multiphase approach: application to a fire spread
model across a fuel bed. Combustion Science and
Technology 162, 59-84.
SIMEONI A., SANTONI P.A., LARINI M., BALBI J.H. 2001.
On the wind advection influence on the fire spread
across a fuel bed: modelling by a semi-physical
approach and testing with experiments. Fire Safety
Journal 36, 491-513.
SIMEONI A., SANTONI P.A., BALBI J.H. 2002. A strategy
to elaborate forest fire spread models for
management tools. International Journal of
Modelling and Simulation 4, 213-224.
SIMEONI A., SANTONI P.A., LARINI M. BALBI J.H. 2003.
Reduction of a multiphase formulation to include a
simplified flow in a semi-physical model of fire
spread across a fuel bed. International Journal of
Thermal Science 42, 95-105.
SÉRO-GUILLAUME O. 2003, Presentation and
comparison of forest fires propagation models,
Workshop on forest fire modelling, invited
conference, State University of Florida, Tallhasee,
pp. 12-14 March.
SERO-GUILLAUME O., J. MARGERIT 2002. Modelling
forest fires. Part 1 : a complete set of equations
derived by extended irreversible thermodynamics.
Int. J. of Heat & Mass Transfer Vol. 45: 1795-1722p.
SERO-GUILLAUME O., MARGERIT J. 2002. Modelling
forest fires. Part 1 : a complete set of equations
derived by extended irreversible thermodynamics.
Int. J. of Heat & Mass Transfer, Vol. 45, pp. 1795-
1722.
SERO-GUILLAUME O., MARGERIT J., BOTELLA O., 2002.
Asymptotic analysis: the way from complete
physical models to propagation models, IV
International Conference on Forest Fires Research.
Coimbra.
STEWARD, F.R. 1971. A mechanistic fire spread model.
Combustion Science and Technology Vol.4:177-
186.
18
EUFIRELAB

TACHAJAPONG W., ZHOU X., MAHALINGAM S. 2006. VIEGAS D.X. 2002. Fire line rotation as a mechanism for
Numerical Modeling of Fire Spread through Trees fire spread on a uniform slope. International Journal
and Shrubs. In Proc. V Int. Conf. on Forest Fire of Wildland Fire 11(1): 11-23.
Research, Viegas, D.X. (Ed.), Figueira da Foz, 27-
VIEGAS D.X., Pita L.P. 2004. Fire Spread in Canyons.
30 Nov. Elsevier Publishers.
International Journal of Wildland Fire 13(3):1-22.
TELISIN, H.P. 1974. Flame radiation as a mechanism of
WEBER, R.O. 1989. Analytical models for fire spread
fire spread in forests. In: Heat Transfer in Flames
due to radiation. Combustion and Flame 78:398-
(Edited by John Wiley and Sons). Chapter 29,
408.
pages 441-449.
WEBER, R.O. 1991a. Toward a comprehensive wildfire
THOMAS, P.H. 1967. Some aspects of the growth and
spread model. International Journal of Wildland Fire
spread of fire in the open. Forestry 40:139-164.
1(4) :245-248.
THOMAS, P.H. 1971. Rates of spread of some wind-
WEBER, R.O. 1991b. Modelling fire spread through fuel
driven fires. Forestry 44:155-175.
beds. Progress in Energy and Combustion Science
VAN WAGNER, C.E. 1967. Calculations on forest fire 17:67-82.
spread by flame radiation. Forestry Branch
Publication No 1185, Canadian Department of
Forestry, 14 pages.

D-03-09.doc 19
EUFIRELAB
3 A REVIEW ON WILDLAND FIRE SMOKE DISPERSION MODELLING
3.1 INTRODUCTION
Biomass burning is:
- a locally, regionally, and globally important
biospheric phenomenon, and
- a significant source of various environmentally
significant gases and solid particulate that can
produce severe degradations of air quality on a local
and/or regional scale.
As a consequence of the increase of frequency and
severity of forest fires, in particular in the wildland-
urban interface (WUI), populations are frequently
exposed to unhealthy concentrations of toxics.
It is worsey when fire occurs in the vicinity of large
urban agglomerations, as recently in Southeast Asia,
Australia, South America, Russia, or Southern Europe.
In fact, there is a growing awareness that smoke
from wildland and prescribed fires can expose
individuals and populations to hazardous air pollutants.
This represents an important public health issue for
the affected communities and the personnel involved in
firefighting operations (BRUSTET et al. 1991, Ward et al.
1993, MIRANDA et al. 1994, 2005, REINHARDT et al.
2001).
According to the literature in the subject (e.g.,
ROTHMAN et al. 1991), fire workers can experience
subchronic, and chronic effects of exposure to forest
fires.
Adverse health effects begin with acute,
instantaneous eye and respiratory irritation and
shortness of breath but can develop into headaches,
dizziness, and nausea lasting up to several hours.
Longer-term health effects, lasting days to perhaps
months, have also been identified among wildland fire-
fighters.
These decrements in lung function include:
- a slightly diminished capacity to breathe,
- constriction of the respiratory tract, and
- hypersensitivity of the small airways.
The spatial scale of the problem gives it an even
larger dimension.
Notwithstanding that even major wildfires are limited
to some hundreds of hectares, their impacts, knowing
no natural or political boundaries, can be felt and
reported far beyond the physical limits reached by fire
spread.
Depending on meteorological conditions, biomass-
burning plumes and haze layers can persist at the
atmosphere for long periods of time influencing the
chemical and optical characteristics of the atmosphere.
NASA mission conducted at South Pacific found
approximately 20 days old smoke plumes (BLAKE et al.
1999):
- travelling up to 1000 km or more (BROWELL et al.
1996) and
- extending horizontally over 100 km as observed
over the central Amazon Basin (ANDREAE et al.
1988).
D-03-09.doc
3.2 EMISSION AND DISPERSION OF AIR
POLLUTANTS
3.2.1 The major air pollutants
Since their formation to final impacts, forest fires-
generated air pollutants undergo in an extremely
complex process:
- it begins with the emission from the biomass
combustion,
- continues through the transport phase, during which
pollutants are transformed and dispersed, and the
subsequent deposition and penetration into the
canopy, and
- ends at the exposed organisms, with consequent
effects on health, especially among individuals from
groups at risk.
Biomass smoke is a complex mixture that contains a
large and diverse number of chemicals, including both
particulates and gaseous compounds.
Following LEVINE (1999), typical major air pollutants
resulting from forest fires are:
- PM: particulate matter
- CO2: carbon dioxide
- CO: carbon monoxide
- CH4: methane
- NMHC: non-methane hydrocarbons
- NO: nitric oxide
- N2O: nitrous oxide, and
- NH3: ammonia
CO, CH4, NMHC, and NO are chemically active
gases that strongly influence the local/regional
concentrations of the major atmospheric oxidants
ozone O3 and the hydroxyl radical OH.
Some measurements suggest that biomass burning
may be a significant global source of methyl bromide
(LEVINE et al. 1995), which leads to the chemical
destruction of O3 in the stratosphere.
Production of aerosols is also very important, giving
rise to local pollution, and affecting the radiation budget
of the Earth and, hence, influencing global climate.
3.2.2 Emission of particulate matter
In some large wildfires, source strengths exceeding
0.6 tons of particles per second can be attained (WADE
and WARD 1973).
PM emission factors range from 2 to 90 g.kg-1,
depending on forest fuel types, fuel arrangement and
combustion characteristics (MAHAFFEY and MILLER
2001).
Although the distribution of PM depends on the
combustion conditions, the available data indicates that
the sizes of most particles produced during a fire can
be described through a bimodal distribution, similarly to
the emissions from any other combustion source.
A fine particle mode can be defined with a maximum
aerodynamic diameter of 2.5 µm and a mean value of
0.3.
20
EUFIRELAB

Additionally, ultra-fine particles, that rapidly 3.2.4 Transportation of the mixture

agglomerate to form fine particles, can be defined as
After formation, the transport of the highly dynamic
those smaller than 0.1 µm.
mixture of combustion compounds away from the
A coarse particle mode is characterized by a mean
emission site depends on the velocity and
diameter greater than 10 µm and agglomerates all the
thermodynamic fields of the atmosphere.
particles larger than 2.5 µm.
The transport patterns are ultimately influenced by
The size distribution of biomass burning smoke has
the interactions of the intense turbulent and convective
been measured by several investigators, indicating that
circulations induced by the heat released with the
in terms of number the majority of particles emitted are
surrounding atmosphere.
ultra-fine, with a peak in the distribution between 0.15
During transport, air pollutants undergo in
and 0.4 µm and only a small fraction in the larger size
physicochemical transformation processes in which the
range (BRAUER 1998, MORAWSKA et al., 1998).
compounds change their chemical composition,
This was evidenced by SANDBERG and MARTIN
physical characteristics and concentration.
(1975) works, which indicated that:
- 82% of the particle mass was constituted by The residence time of combustion compounds in air:
particles smaller than 1 µm and - depends on the nature of these processes and,
- 69% smaller than 0.3 µm. depending on the size range of particles and
This was also evidenced by the ground-based meteorological conditions,
particles samplings conducted by WARD and HARDY - varies from seconds or minutes for very large
(1989), which showed the bimodal distribution with only particles, to days or weeks in particles smaller than
a small fraction of the total mass (less than 10%) 1 µm.
between 2 and 10 µm.
This fact was reported by MORAWSKA et al. (1998),
3.2.3 Modification of the mixture who observed that at a distance of 20 km from a fire
there is a significant increase only in the concentration
The overall mixture of combustion products is also
of particles with less than 1 µm, indicating that, while
affected by the combustion phase.
travelling over this distance, coarser particles have
In fact, the smouldering phase, known to be a very
been removed from the atmosphere.
low-intensity combustion process, releases several
In fact, due to high gravitational settling velocities of
times more fine particles than flaming combustion.
large particles, only fine particles can remain
In this last, the percentage of fine particles produced
suspended in the atmosphere the necessary time to
ranges from 80 to 95%, depending, among other
permit their transportation over long distances.
factors, on the turbulence intensity in the combustion
Also smaller particles are transformed, mainly by
zone (WARD 1999).
coagulation and condensation into larger particles, and
Moreover, also fuel properties can induce a
are subsequently removed through gravitational settling
substantial effect on the percentage of vegetation
or by in-cloud scavenging during precipitation.
consumed by smouldering combustion:
- from 10% in savannah ecosystems The time of day or night is of extreme importance in
- to 90% in peat, rotten logs or rotten wood residues the transportation of smoke plumes.
(WHO/UNEP/WMO, 1999). In the presence of nocturnal jets with speeds of
25 m.s-1 smaller particles can be transported up to
Also the addition of fire retardant chemicals can
1000 km during a 10-hour period (GARSTANG 1999).
affect the type and quantity of air pollutants emitted.
In fact, an increase on the amount of particles
produced was already reported as a consequence of
the diminishing of the pyrolysis temperature (GEORGE
and SUSOTT 1971, PHILPOT et al. 1972, Kalabokidis
2000).
Fire retardant chemical formulation is also known to
affect the amount of smoke produced.
In fact, ammonium phosphate retardants reduce
glowing combustion and are much more effective in
decreasing the rate of weight loss and increasing
residue and smoke production than ammonium
sulphate retardants (GEORGE and SUSOTT 1971,
PHILPOT et al. 1972).

D-03-09.doc 21
EUFIRELAB
3.3 NUMERICAL MODELLING OF SMOKE
PRODUCTION AND DISPERSION
The study of the effect of forest fire emissions on air
quality requires an integrated approach of different
components, such as emissions, atmospheric flow, fire
progression, pollutants dispersion and associated
impacts, as visibility reduction or human health effects.
The numerical modelling of this entire complex
process, which is intended to provide useful information
for managing or interpreting fire effects, is not,
however, a trivial task.
The knowledge of fire characteristics and the spatial
and temporal distribution of products originated by a
fire require an overall understanding of all the
atmospheric processes involved, which can have
extremely distinct spatial scales:
- from motions smaller than those of the fire itself,
- to circulations occupying a significant fraction of the
planet (GARSTANG 1999).
To estimate the effects of smoke on air quality the
emission production calculation is insufficient.
Beyond the knowledge of the emission
characteristics, air pollution assessment requires the
calculation of the concentration of a given pollutant
some distance from its source, taking into account all
the phenomena that affect transport and dispersion of
smoke.
In practice, this procedure can be considered to
have three main components, which are not
necessarily independent of each other:
- to estimate the plume rise,
- to calculate the transport speed and direction of the
emitted pollutants, and finally,
- to simulate smoke dispersion.
3.3.1 To estimate the plume rise
Fire emission products vary greatly with the type of
fuel consumed and its moisture content, the fire
intensity or rate of energy released, and other factors
such as humidity and wind speed.
Therefore, in order to estimate the emission rate of
a pollutant, variables as the fuel load, combustion rate,
and the emission factor are needed.
Fire, as an intense heat source, induces the
formation of heat-driven turbulence and convective
motions, which interact with the atmospheric flow field
according to a complex and non-linear process.
The heat released originates the formation of a
convection column that forces the emitted pollutants to
ascend.
Upward velocities within the fire can range from 20
to 40 m.s-1, determining the height attained by the
centre of the smoke plume.
In fact, there is a considerable uncertainty on the
knowledge of the mean height reached by fire products,
which is a critical factor to determinate the distance
through which these combustion products will be
transported (GARSTANG 1999).
D-03-09.doc
With the diminishing of the rate of energy released
the plume looses its columnar shape to a point where
lift of emissions results mainly from vertical
atmospheric mixing.
This fact demonstrates that plume rise is a function
of not only heat release rate, but also atmospheric
stability and wind speed.
Therefore, its modelling requires to calculate the
rates of heat release and emission for each gas
species and particle size, and also to know the local
meteorological parameters (temperature, wind and
humidity) and mixing height of the atmosphere.
PHURO et al. (1976) presented a set of equations to
calculate the plume height, based on the plume rise
relations developed by BRIGGS (1969, 1971 and 1972)
for stack emissions and in the stability classes of
PASQUILL (1975)
Over the years, different procedures and tools have
been developed to:
- estimate fuel loading and consumption, and
- characterize the resulting emissions from prescribed
burns and wildfires.
Three models should be emphasized within this
global research effort (SANDBERG et al. 1999, SESTAK et
al. 2002):
- the fuel consumption models FOFEM First Order
Fire Effects Model (REINHARDT et al. 1997)
- CONSUME (OTTMAR et al. 1993a and 2000) and
- the source strength model EPM Emissions
Production Model (SANDBERG and PETERSON 1984),
They are used to provide emission and heat release
data for most dispersion models.
A number of other models should also be referred,
namely:
- SMSINFO (OTTMAR et al. 1993b);
- FETM Fire Emissions Tradeoff Model (SCHAAF et al.
1996);
- ACOST Automatic Calculation of Slash Tonnage
model and PCOST Pile Tonnage Calculation
Worksheet (ACOST/PCOST 2000); and
- FASTRACS Fuel Analysis, Smoke Tracking, and
Report Access Computer System (url 1).
3.3.2 To calculate the transport speed and
direction
The vertical distribution of local air temperature and
wind fields governs the recirculation of fire-generated
products.
Once the smoke plume has reached its maximum
rise height, it becomes completely dominated by the
local wind field.
Thus, the accurate determination of the three-
dimensional wind field is of fundamental importance,
especially in areas of complex topography where
mountain-valley or slope-valley winds are expected to
occur.
Factors that affect transport and dispersion are, in
that stage, atmospheric stability, temperature
inversions, mixing height, wind speed and direction and
local-scale systems as land and sea breezes (PHURO et
al. 1976).
22
EUFIRELAB
In very simple models, the trajectory and dispersion
of the smoke plume can be simulated through straight-
line trajectories.
Others, assuming that the atmosphere is neutrally
buoyant, compute the trajectories of air parcels that are
transported by the three-dimensional wind field, which
is calculated by a numerical weather prediction NWP
model.
This so-called “trajectory technique” is considered to
be simple in its concept, and requiring modest
computer resources.
Trajectories can be run:
- forward in time to determine receptor areas, or
- backwards in order to determine the pollutant
source areas.
In general, multiple trajectories are required due to
the instability of the atmospheric flow.
Another technique used to simulate the behaviour of
the smoke plume is to apply an atmospheric transport
model ATM, based on the conservation of mass for the
considered pollutant.
Its movement through the atmosphere, resulting
from the mean wind field provided by the NWP model
and the turbulent mixing processes (parameterised in
the ATM), is balanced by the difference between the
emission inputs and pollutant losses by wet and dry
deposition (also parameterised in the ATM).
Two main modelling approaches are applied in this
kind of numerical tools:
- either Lagrangian models that follow the trajectories
of segments, puffs or particles; or
- Eulerian models, which solve the diffusion equation
at every point on a fixed grid.
In Lagrangian models, the dispersion is
accomplished by Gaussian (or an equivalent probability
function) diffusion for segments and puffs and by Monte
Carlo (Langevin-Markov) techniques for particles.
For Eulerian models, the dispersion is usually
performed by first order turbulent closure, but higher-
order turbulent closures could be used (TAPPER and
HESS 1999).
3.3.3 To simulate smoke dispersion
According to the type of application, smoke
dispersion models can be classified as follow
(BREYFOGLE and FERGUSON 1996, MILLER 2001):
Research models
They are capable to model many chemical
compounds, concentrations, and a large radius of
receptor sites and their concentrations.
Regulatory/Planning models
They could be used for permit approval and would
be capable to show dispersion over large geographic
areas.
They offer the possibility to run various weather
scenarios in advance of the actual fire.
However, because any model has been yet officially
validated for biomass burning, regulatory use would be
inappropriate at this time.
D-03-09.doc
Screening models
They are use by field users concerned with smaller
prescribed burns.
They would aid to visualise what fuel and weather
conditions are best suited.
They could be used as a planning tool before fire
and at the time of the fire with the real-time
meteorological data.
Smoke dispersion models are becoming
increasingly valuable tools in smoke management,
especially for screening and planning.
The expected result of all models is the ability to
estimate the variation in time and space of particle and
gas concentrations that affect human health and alter
visibility.
Several models have been developed for the
simulation of smoke transport and dispersion
(BREYFOGLE and FERGUSON 1996, MIRANDA 1999),
namely:
- SASEM Simple Approach Smoke Estimation Model
(SESTAK and RIEBAU 1988),
- HYSPLIT Hybrid Single-Particle Langrangian
Integrated Trajectory model (DRAXLER 1992,
DRAXLER and HESS 1997),
- TSARS Plus (HUMMEL and RAFSNIDER 1995),
- CALPUFF (SCIRE et al. 1995 and 2000),
- AIRFIRE (MIRANDA 1998 and 2004),
- DISPERFIRE (MIRANDA 1998),
- VALBOX Ventilated Valley Box model (SESTAK et al.
1988),
- TAPAS (FOX et al. 1987),
- VSMOKE (LAVDAS 1996),
- VSMOKE-GIS (HARMS and LAVDAS 1997),
- NFSPUFF (HARRISON 1996),
- REMSAD Regional Modelling System for Aerosols
and Deposition (ICF Consulting 2002), which is
based on the variable-grid UAM-V Urban Airshed
Model,
- a regional-scale photochemical modelling system
(SAI 1999), or
- CMAQ Community Multi-scale Air Quality modelling
system (USEPA 1999).
Some of the referred models integrate the
determination of the three-dimensional wind field
through the interpolation of surface wind
measurements.
In this case, the interpolation becomes more
accurate as the number of input observation sites
increase.
Others, on the contrary, assume a constant value of
wind velocity and direction, which can be considered as
a very rude approach, since wind fields occurring in
nature are far from being uniform, especially in areas of
complex topography.
23
EUFIRELAB
In the generation of the necessary meteorological
fields, the coupling with meteorological models
assumes a particular importance, both
- local, as is the case of NUATMOS (ROSS et al.
1988), or
- mesoscale ones, as MM5 (DUDHIA, 1993), MEMO
(FLASSAK and MOUSSIOPOULOS 1987) or
- RAMS (PIELKE et al. 1992).
Mesoscale circulation can be very important
because they:
- directly affect the dispersion of smoke by causing
abrupt changes in local atmospheric stability and
- can influence the direction in which smoke will be
transported.
Despite the large number of smoke dispersion
models, some management systems include the
efficiency of the atmospheric dispersion of smoke using
dispersion indexes.
An example of this type of tools is the FFMIS Florida
Fire Management Information System (BRENNER et al.
1997).
It is a fire management system that simulates the
current and forecast state of the:
- surface weather,
- atmospheric stability,
- fire danger,
- fire weather index and
- fire behaviour potential.
In this system, the efficiency of the atmospheric
dispersion of smoke is estimated using the ADI
Atmospheric Dispersion Index, which indicates the
efficiency of the atmosphere at carrying gaseous or
small PM away from its source.
ADI is based on the Gaussian Plume statistical
model (LAVDAS 1986), considering a Gaussian
distribution of the pollutant concentration in a finite box
downwind of the source.
In terms of visibility modelling, the optical effects in
the atmosphere due to air pollution can be adequately
described by examination of the atmospheric light
extinction.
This approach assumes that the increased
extinction due to an increment in a specific air pollutant
species is directly proportional to the increase in mass
of the species.
Two basic approaches have been used to establish
the relationships between atmospheric composition and
optical properties that dictate the visibility impairment.
One is based in the estimation of light extinction
using extinction efficiencies derived theoretically, while
in the other extinction efficiencies are obtained by
empirical methods.
The most reliable estimates of extinction due to
individual chemical species are those computed using
theoretical light scattering models (Mie-theory) fitted to
the size-resolved composition of the aerosols observed
in the region of interest.
Such models explicitly simulate the physical cause-
and-effect relationship and best utilize all of the data on
the aerosol's properties that affect scattering
efficiencies (TRIJONIS et al. 1991).
D-03-09.doc
However, even the most fundamental of these
models incorporate important assumptions concerning
particle structure, condensed water, and other
unobserved aspects of the aerosol.
Many researchers have developed versions of Mie
scattering models (GRAY and KLEINHESSELING 1996).
A significant limitation to the use of theoretically
derived extinction efficiencies is that they have not
been developed and tested for many locations.
A much wider range of extinction efficiency
estimates is available from multiple regression
analyses of the empirical relationship between
extinction and aerosol composition (TRIJONIS et al.
1991).
In this approach, the observed light scattering is
statistically compared to observed aerosol
concentration and composition to estimate the
scattering efficiencies of each particulate species.
The hygroscopic species (e.g., sulphate and nitrate)
concentrations are usually corrected to account for
water uptake using a function of the ambient relative
humidity.
This approach has generated statistically significant
extinction relationships for many applications, and the
resulting coefficients have been used to accurately
estimate light scattering from aerosol concentration
data (GRAY and KLEINHESSELING 1996).
The computational tools used to estimate visibility
from aerosol concentrations and compositions are often
found as post processing modules attached to air
quality models.
Various modelling approaches that have been used
to relate air quality to visibility, namely:
- VASM / ASTRAP (DOE 1994),
- MESOPUFF II (USEPA, 1993),
- CALPUFF (SCIRE et al. 1995),
- NPAQMS (RIM II) (GRAY et al. 1993),
- VISCREEN / PLUVUE II (USEPA 1992),
- STAGHAZE (LATIMER 1993),
- RIVAD (LATIMER 1990),
- GCVTC (USEPA 1995),
- ELSIE (SLOANE et al. 1991) and
- ROME (GABRUCK et al. 1999).
A third approach consists in developing a composite
set of extinction efficiencies from a variety of empirical
and theoretical analyses.
Empirically derived extinction efficiencies obtained
using multiple linear regression models are subject to a
number of uncertainties.
Errors in the resulting extinction efficiencies can be
related to errors in light scattering (bsp) measurements,
correlations between chemical concentrations, and
non-linearities caused by variations in size distribution
and mixing state from sample to sample ‘GRAY and
KLEINHESSELING 1996).
In the United States, for the assessment of values of
air quality and visibility at risk from wildland fires, a 40-
year database was generated, providing the first
nationally consistent map of surface wind and
ventilation index, which represents the product of wind
speed and mixing height.
24
EUFIRELAB
An interactive VCIS ventilation climate information
system was produced (FERGUSON et al. 2003),
allowing users to assess risk based on frequency
patterns of poor, marginal, fair, and good ventilation
conditions through internet (url 2).
In spite of the number of available models to
simulate smoke dispersion and although the existence
of some systems already covering the main questions
to be taken into account, a lack of integration
concerning fire progression stills remains.
In general, the wind field distribution, the fire line
advance, and the interaction between fire and the
ambient wind are not taken into account.
In this context the Universities of Coimbra and
Aveiro, in Portugal, produced the fire behaviour system
DisperFireStation (VALENTE et al. 2007), which was
developed to estimate:
- fire progression,
- smoke dispersion and
- visibility impairment at a local scale.
This system results from the improvement and
integration of two already available numerical tools,
DISPERFIRE (MIRANDA et al. 1994) and FireStation
(LOPES et al., 2002).
FireStation is a software system aimed at the
simulation of fire spread over complex topography.
DISPERFIRE is a real time system developed to
simulate the dispersion in the atmosphere of the
pollutants emitted during a forest fire.
In addition, a model for the estimation of visibility
impairment, based on the relationship between the air
pollutants concentration and visibility was included.
The whole system was developed under a graphical
interface, previously created for FireStation, allowing a
friendly user access and providing easily readable
output to facilitate its application under operational
conditions.
A detailed description of the system can be found in
Deliverable D-03-04.
Figure 1 presents an example of an output of the
model obtained from the simulation of the flow field and
the CO dispersion in an area centred in Gestosa,
during the 2004 fire experiments.
The simulation domain is, in this case about 2 km2.
Current fire assessment systems are starting to be
planned in order to integrate all the needed variables in
a common and user-friendly tool.
In the scope of the development and implementation
of advanced methods for high-resolution fire and
smoke modelling, USDA-FS U.S. Department of
Agriculture Forest Service recently established the
FCAMMS Fire Consortia for Advanced Modelling of
Meteorology and Smoke (url 3), involving all the USDA-
FS Research Stations and their partners, as
represented in figure 2.
D-03-09.doc
Under this collaborative and coordinated national
effort to model smoke impacts, BlueSky Smoke
Modelling Consortium is specially devoted to the
development and application of real-time smoke
modelling to support fire operations and smoke
management.
This automated smoke modelling framework,
designed to rely upon real-time meteorological forecast
data and existing air quality models, tracks daily
emissions and predicts the cumulative concentrations
of smoke from prescribed fires and wildfires (FERGUSON
et al. 2001, SESTAK et al. 2002).
In BlueSky, fire emission estimates are obtained
through the coupling of CONSUME and EPM with real-
time fire activity reports.
Real-time estimates and forecasts of smoke impacts
for active smoke management programs are generated
through the integration of the high-resolution
mesoscale meteorological model MM5 with EPM and
the dispersion model CALPUFF.
In the near future, it is expected that the BlueSky
framework will also be running the CMAQ Eulerian grid
model.
Another effort in the development of appropriate
emissions models for wildland burning is CSEM
Community Smoke Emissions Model; it:
- which couples CONSUME with EPM
- utilises national fuels coverage in GIS format, and
- is specifically designed to provide historical fire
emissions estimates for use in CMAQ and
REMSAD.
Smoke forecasts are produced by the NOAA-ARL
National Oceanic and Atmospheric Administration's Air
Resources Laboratory using HYSPLIT HYbrid Single-
Particle Lagrangian Integrated Trajectory model, which
is the newest version of a complete system for
computing simple air parcel trajectories to complex
dispersion and deposition simulations.
Forecasts are available in url 4, and an example is
shown in figure 3.
This tool is integrated in the web-based system
called READY Real-time Environmental Applications
and Display sYstem, developed for:
- accessing and displaying meteorological data and
- running trajectory and dispersion model products on
the web server.
This system brings together dispersion models,
graphical display programs and textual forecast
programs.
WFAS Wildland Fire Assessment System, which
was first made available in 1994 and is still under
development.
Although it does not simulate the dispersion of
smoke, it constitutes a valuable tool for modellers as an
internet-based (url 5) integrated information system
planned to provide information on weather and fire
potential for the entire territory of the U.S.
An example is shown in figure 4.
25
EUFIRELAB
INTERFACE Project, funded by the Portuguese
National Foundation for Science and Technology,
addresses the assessment of the effects of WUI fires
on air, meeting the growing need for information related
to smoke impacts management.
The proposed work involves the development and
application of an air quality modelling system for the
simulation and forecast of smoke dispersion on an
integrated mesoscale / microscale basis.
An innovative and substantial aspect is the
integration of all the steps involved within this global
process:
- fire behaviour description;
- emission modelling systems;
- plume transport,
- dispersion,
- reaction and deposition mechanisms; and
- visibility impairment.
D-03-09.doc
3.4 FINAL COMMENTS
Depending on the nature of transport, dispersion
and removal processes, the residence time of the
emitted combustion compounds is highly variable,
varying from few seconds to some weeks.
The study of the changes in chemical composition,
physical characteristics and atmospheric concentration
in which this highly dynamic mixture undergoes
requires an integrated approach of all the involved
components:
- fire progression and emissions, atmospheric flow
and
- air pollutants dispersion, reaction and deposition.
However, a lack of integration between all the
components of the process, since emission to health
effects and visibility reduction, still compromises the
accuracy of the result.
The development of powerful and accurate codes
for the numerical simulation of the involved phenomena
still represents a challenge to science.
26
EUFIRELAB
3.5 REFERENCES
ACOST/PCOST. 2000. Oregon Department of
Forestry, Fuels and Meteorological Services
Section. Available in: http://www.odf.state.or.us.
ANDREAE, M., BROWELL, E.V., GARSTANG, M., GREGORY,
G.L., HARRISS, R.C., HILL, G.F., JACOB, D.J.,
PEREIRA, M.C., SACHSE, G.W., SETZER, A.W., SILVA
DIAS, P.L., TALBOT, R.W., TORRES, A.L. and WOFSY,
S.C. 1988. Biomass-Burning Emissions and
Associated Haze Layers Over Amazonia, Journal of
Geophysical Research, 93(D2). pp 1509–1527.
BLAKE, N.J., BLAKE, D.R., WINGENTER, O.W., SIVE, B.C.,
MCKENZIE, L.M., LOPEZ, J.P., SIMPSON, L.J.,
FUELBERG, H.E., SACHSE, G.W., ANDERSON, B.E.,
GREGORY, G.L., CARROLL, M.A., ALBERCOCK, G.M.
and ROWLAND, F.S. 1999. Influence of Southern
Hemispheric Biomass Burning on Mid Tropospheric
Distributions of Nonmethane Hydrocarbons and
Selected Halocarbons Over the Remote South
Pacific. Journal of Geophysical Research, 104. pp.
16213-16232.
BRAUER, M. 1998. Health Impacts of Biomass Air
Pollution. Report prepared for the Biregional
Workshop on Health Impacts of Haze-related Air
Pollution, 1-4 June, Kuala Lumpur, Malaysia.
BRENNER, J., ARVANITIS, L., BRACKETT, D., LEE, B.,
CARR, R. and SUDDABY, R. 1997. Integrating GIS,
Meso-Scale Fire Weather Prediction, Smoke Plume
Dispersion Modeling, and the Internet for Enhanced
Open Burning Authorizations and Wildfire Response
in Florida. In: Proceedings from the 1997 ESRI
International Users Conference. 8-11 July, San
Diego, California, USA.
BREYFOGLE, S. and FERGUSON, S.A. 1996. User
Assessment of Smoke-Dispersion Models for
Wildland Biomass Burning. Gen. Tech. Rep. PNW-
GTR-379. U.S. Department of Agriculture, Forest
Service, Pacific Northwest Research Station.
Portland, OR, EUA. 30 p.
BRIGGS, G.A. 1969. Plume Rise. Air Resour Atmos.
Turbul. and Diffus. Lab., Environ. Sci. Serv. Adm.,
U.S. Atomic Energy Comm., Oak Ridge, Tenn.,
USA. 81 p.
BRIGGS, G.A. 1971. Some Recent Analyses of Plume
Rise Observations. In: Proceedings from the 2nd Int.
Clean Air Congr., Acad. Press. New York, USA. pp.
1029-1032.
BRIGGS, G.A. 1972. Discussion on Chimney Plumes in
Neutral and Stable Surroundings. Atmos. Environ.
6(7). pp: 507-510.
D-03-09.doc
BROWELL, E.V., FENN, M.A., BUTLER, C.F., GRANT,
W.B., CLAYTIN, M.B., FISHMAN, J., BACHHMEIER, A.S.,
ANDERSON, B.E., GREGORY, G.L., FUELBERG, H.E.,
BRADSHAW, J.D., SANDHOLM, S.T., BLAKE, D.R.,
HEIKES, B.G., SACHSE, G.W., SINGH, H.B. and
TALBOT, R.W. 1996. Ozone and Aerosol
Distributions and Air Mass Characteristics Over the
South Atlantic Basin During Burning Season.
Journal of Geophysical Research, 101. pp. 24 043-
24 068.
BRUSTET, J., VICKOS, J., FONTAN, J., PODAIRE, A. and
LAVENU, F. 1991. Characterisation of active fires in
West African savannahs by analysis of satellite
data: Landsat thematic mapper. In: LEVINE, J. (Ed.),
Global biomass burning: atmospheric, climatic, and
biospheric implications. The Massachussets
Institute of Technology Press, Cambridge,
Massachussets, London, pp. 53-60.
DOE. 1994. VASM Q's and A's: An Overview of the
Visibility Assessment Scoping Model (VASM) and
Model Set. Quick Reference Fact Sheet, U.S.
Department of Energy, Washington, D.C., USA.
DRAXLER, R.R. 1992. HYbrid Single-Particle
Langrangian Integrated Trajectories (HY-SPLIT):
version 3.0 - User’s Guide and Model Description.
NOAA Tech. Memo. ERL ARL-195.
DRAXLER, R.R. and HESS, G.D.1997. Description of the
HYSPLIT-4 Modeling System. NOAA Technical
Memorandum ERL ARL-224.
DUDHIA, J. 1993. A Nonhydrostatic Version of the Penn
State – NCAR Mesoscale Model: Validation Tests
and Simulation of an Atlantic Cyclone and Cold
Front. Mon. Wea. Rev., 121. pp. 1493-1513.
FERGUSON, S.A., MCKAY, S.J., NAGEL, D.E., PIEPHO, T.,
RORIG, M.L., ANDERSON, C. and KELLOGG, L. 2003.
Assessing Values of Air Quality and Visibility at Risk
from Wildland Fires. Res. Pap. PNW-RP-550. U.S.
Department of Agriculture, Forest Service, Pacific
Northwest Research Station. Portland, OR, USA. 59
p.
FERGUSON, S.A., PETERSON, J. and ACHESON, A. 2001.
Automated, Real-Time Predictions of Cumulative
Smoke Impacts From Prescribed Forest and
Agricultural Fires. In: Proceedings from the 4th
Symposium on Fire and Forest Meteorology, 13-15
November 2001, Reno, Nevada, American
Meteorological Society. pp. 168-175.
FLASSAK, T. and MOUSSIOPOULOS, N. 1987. An
Application of an Efficient Non-Hydrostatic
Mesoscale Model. Boundary Layer Meteorology,
Vol. 41. pp. 135-147.
FOX, D., ROSS, D., DIETRICH, D., MUSSARD, D. and
RIEBAU, A. 1987. An Update on TAPAS and its
Model Components. Proceedings of the Ninth
Conference on Fire and Forest Meteorology. April,
21-24, San Diego, California. pp. 1-7.
27
EUFIRELAB
GABRUCK, R.S., SYKES, R.I., SEIGNEUR C., PAI P.,
GILLESPIE P., BERGSTROM, R.W. and SAXENA, P.
1999. Evaluation of the Reactive and Optics Model
of Emissions (ROME). Atmospheric Environment,
Vol. 33. pp. 383 -399.
GARSTANG, M. 1999. The Role of the Atmosphere in
Fire Occurrence and the Dispersion of Fire
Products. In: Health Guidelines for Vegetation Fire
Events - Background Papers. Kee-Tai-Goh,
Schwela D., Goldammer J.G. and Simpson O.
(Eds). United Nations Environment Programme,
Nairobi, World Health Organization, Geneva, World
Meteorological Organization, Geneva, Institute of
Environmental Epidemiology, WHO Collaborating
Centre for Environmental Epidemiology, Ministry of
the Environment, Singapore. pp 101-121.
GEORGE, C.W. and SUSOTT, R.A. 1971. Effects of
Ammonium Phosphate and Sulfate on the Pyrolysis
and Combustion of Cellulose. Research Paper INT-
90. U.S. Department of Agriculture, Forest Service.
27 p.
GRAY, H.A. and KLEINHESSELING, D. 1996. Evaluation of
Existing Information of the Effects of Air Pollutants
on Visibility in the Southern Appalachians. Systems
Applications International.
GRAY, H.A., STOECKENIUS, T.E., EMERY, C.A. and
RICHARDS L.W. (1993) Assessment of the Effects of
the 1990 Clean Air Act Amendments on Visibility in
Class I Areas. Systems Applications International,
San Rafael, California (SYSAPP-93/162).
HARMS, M.F. and LAVDAS, L.G. 1997. Users Guide to
VSMOKE-GIS for Workstations. Research Paper
SRS-6. U.S. Department of Agriculture, Forest
Service, Southern Research Station. Asheville, NC,
USA. 41 p.
HARRISON, H. 1996. A User’s Guide to “NFSPUFF”: A
Dispersion Model for Smoke Management in
Complex Terrain. Seattle, WA. A final report
available from Air Program Manager, USDA For.
Ser., Northern Region, Missoula, MT. 42 pp.
HUMMEL, J. and RAFSNIDER, J. 1995. TSARS Plus
Smoke Production and Dispersion Model User's
Guide. Preliminary Draft 7.95. National Biological
Service and the Interior Fire Coordination
Committee. 96 p.
ICF CONSULTING. 2002. User’s Guide to the Regional
Modeling System for Aerosols and Deposition
(REMSAD). Version 7. July 2002. 153 pp.
KALABOKIDIS, K.D. 2000. Effects of Wildfire Suppression
Chemicals on People and the Environment - A
Review. In: Global Nest, the Int. J. Vol. 2, No 2,
Greece. pp 129-137.
LATIMER, D.A. 1990. Visibility Modeling: The State of
the Technology and New Directions. In: Visibility
and Fine Particles. Mathai, C.V. (Ed.), Air and
Waste Management Association, Pittsburgh,
Pennsylvania.
D-03-09.doc
LATIMER, D.A. 1993. Development of Regional Haze
Screening Models. Air and Waste Management
Association 86th Annual Meeting and Exhibition,
Denver, Colorado (93-TP- 49.04).
LAVDAS, L.G. 1986. An Atmospheric Dispersion Index
for Prescribed Burning. Research Paper SE-256,
US Department of Agriculture, Macon, Georgia,
USA.
LAVDAS, L.G. 1996. Program VSMOKE - Users Manual.
Gen. Tech. Rep. SRS-6. U.S. Department of
Agriculture, Forest Service, Southern Research
Station. Asheville, NC, USA. 147 p.
LEVINE, J., COFER, W., CAHOON, D. and WINSTEAD, E.
Biomass Burning: A Driver for Global Change,
Environmental Science and Technology, 29, 3, pp.
120A-125A, 1995.
LEVINE, J.S. 1999. Gaseous and Particulate Emissions
Released to the Atmosphere From Vegetation Fires.
In: Health Guidelines for Vegetation Fire Events -
Background Papers. Kee-Tai-Goh, Schwela D.,
Goldammer J.G. and Simpson O. (Eds): United
Nations Environment Programme, Nairobi, World
Health Organization, Geneva, World Meteorological
Organization, Geneva, Institute of Environmental
Epidemiology, WHO Collaborating Centre for
Environmental Epidemiology, Ministry of the
Environment, Singapore. pp 280-294.
LOPES, A.M.G., CRUZ, M.G. and VIEGAS, D.X. 2002.
FireStation – an integrated software system for the
numerical simulation of fire spread on complex
topography. Environmental Modelling & Software,
17, 269-285.
MAHAFFEY, L. and MILLER, M. 2001. Air Quality. In: Fire
Effects Guide, National Wildfire Coordinating Group,
Fire Use Working Team. Chapter IV. Available in:
http://fire.r9.fws.gov/ifcc/monitor/FEG.pdf.
MILLER, M. 2001. Computer Software. In: Fire Effects
Guide, National Wildfire Coordinating Group, Fire
Use Working Team. Chapter XII. Available in:
http://fire.r9.fws.gov/ifcc/monitor/FEG.pdf.
MIRANDA, A. I. 1998. Efeito dos Incêndios Florestais na
Qualidade do Ar. PhD. Thesis. Department of
Environment and Planning, University of Aveiro,
Portugal. 213 pp.
MIRANDA, A.I. 1999. Review Commenting on the State-
of-the-Art on Smoke Dispersion. DELFI knowledge
data base. 8 p.
MIRANDA A.I. 2004. An integrated numerical system to
estimate air quality effects of forest fires. Int. J.
Wildland Fire 13(2), pp. 217-226.
MIRANDA, A.I., BORREGO, C., and VIEGAS, D. 1994.
Forest fire effects on the air quality. In ‘Air Pollution
II, Computer Simulation.’(Eds JM Baldasano, CA
Brebbia, H Power, P Zannetti) pp. 191-199.
(Computational Mechanics Publications:
Southampton).
28
EUFIRELAB
MIRANDA, A.I., MARTINS, H., CARVALHO, A. and
BORREGO, C., 2005. Modelling Smoke Effects on
Lisbon Air Quality. In Sixth Fire and Forest
Meteorology Symposium and the19th Interior West
Fire Council Meeting, 25-27 October 2005
Canmore, AB, Canada. Proceedings in CD-Rom.
MORAWSKA, L., THOMAS, S., BOFINGER, N.D.,
WAINWRIGHT, D. and NEALE, D. 1998.
Comprehensive Characterisation of Aerosols in a
Subtropical Urban Atmosphere: Particle Size
Distribution and Correlation with Gaseous
Pollutants. Atmospheric Environment, 32. pp. 2461-
2478.
OTTMAR, R.D., ANDERSON, G.K., DEHERRERA, P.J. and
REINHARDT, T.E. 2000. CONSUME User’s Guide,
Version 2.1. USDA Forest Service Pacific Northwest
Research Station, Fire and Environmental Research
Applications Group, Seattle, Washington, USA.
OTTMAR, R.D., BURNS, M.F., HALL, J.N. and HANSON,
A.D. 1993a. CONSUME Users Guide. Gen. Tech.
Rep. PNW-GTR-304. U.S. Department of
Agriculture, Forest Service, Pacific Northwest
Research Station. Portland, OR, USA. 118 p.
OTTMAR, R.D., BURNS, M.F., TEESDALE, D.R. and HALL,
J.N. 1993b. SMSINFO Users Guide. USDA Forest
Service, Pacific Northwest Research Station, Fire
and Environmental Research Applications Group,
Seattle, Washington, USA.
PASQUILL, E. 1975. Atmospheric Diffusion. 2nd ed.
John Wiley and Sons, New York. 368 p.
PHILPOT, C.W., GEORGE, C.W., BLAKELY, A.D.,
JOHNSON, G.M. and WALLACE JR., W.K. 1972. The
Effect of Two Flame Retardants on Particulate and
Residue Production. Research Paper INT-117.
Intermountain Forest and Range Experiment
Station, Forest Service, U.S. Department of
Agriculture, Ogden, Utah, USA.
PHURO, J.A., LAVDAS, L.G. and BAILEY, P.M. 1976.
Smoke Transport and Dispersion. In: Southern
Forestry Smoke Management Guidebook, Chapter
V, Gen. Tech. Rep. SE-10, U.S. Department of
Agriculture, Forest Service, Southeastern Forest
Experiment Station. Ashville, NC, USA. 140 p.
PIELKE, R.A., COTTON, W.R. and WALKO, R.L. 1992. A
Comprehensive Meteorological Modeling System -
RAMS. Meteor. Atmos. Phys., 49. pp. 69-91.
REINHARDT, E., KEANE, R. and BROWN, J. 1997. First
Order Fire Effects Model: FOFEM 4.0, User's Guide.
General Technical Report INT-344, USDA Forest
Service, Intermountain Research Station, Ogden,
UT.
REINHARDT, E., OTTMAR, R. and CASTILLA, C. 2001.
Smoke Impacts from Agricultural Burning in a Rural
Brazilian Town. Journal of the Air & Waste
Management Association 51, 443-450.
D-03-09.doc
ROSS, D., SMITH, I., MANINS, P. and FOX, D. 1988.
Diagnostic Wind Field Modelling for Complex
Terrain: Model Development and Testing. Journal of
Applied Meteorology, Vol. 27. pp. 785-796.
ROTHMAN N., FORD, D.P., BASER, M.E., HANSEN, J.A.,
O'TOOLE, T., TOCKMAN, M.S. and STRICKLAND, P.T.
1991. Pulmonary function and respiratory symptoms
in wildland firefighters. Journal of Occupational
Medicine 33(11): 1163-1169.
SAI. 1999. User’s Guide to the Variable-Grid Urban
Airshed Model (UAM-V). Systems Applications
International, San Rafael, California.
SANDBERG, D. and MARTIN, R. 1975. Particle Sizes in
Slash Fire Smoke. U.S. Department of Agriculture,
Forest Service, Pacific Northwest Research Station.
Portland, OR, USA.
SANDBERG, D.V. and PETERSON, J.L. 1984. A Source
Strength Model for Prescribed Fires in Coniferous
Logging Slash. Annual Meeting, Air Pollution Control
Association, Pacific Northwest Section. Reprint
#84.20. U.S. Department of Agriculture, Forest
Service, Pacific Northwest Research Station.
Portland, OR, USA. 10 p.
SANDBERG, D.V., HARDY, C.C., OTTMAR, R.D., SNELL,
J.A.K., ACHESON, A.L., PETERSON, J.L., SEAMON, P.,
LAHM, P. and WADE, D. 1999. National Strategic
Plan: Modeling and Data Systems for Wildland Fire
and Air Quality. Gen. Tech. Rep. PNW-GTR-450.
U.S. Department of Agriculture, Forest Service,
Pacific Northwest Research Station. Portland, OR,
USA. 60 p.
SCHAAF, M., SNELL, K., Carlton, D., OTTMAR, R.,
WIITALA, M., HILBRUNER, M. and NESBITT, J. 1996.
Development of the Fire Emissions Tradeoff Model
(FETM) and Application to the Grande Ronde River
Basin, Oregon. Technical Report prepared for
USDA For. Ser., PNW Reg., Contract Nr. 53-83FT-
03-2.
SCIRE, J., STRIMAITIS, D.G., YAMARTINO, R.J. and
XIAOMONG, Z. 1995. A User's Guide for CALPUFF
Dispersion Model. Doc. 1321-2. Concord, MA:
Sigma Research/Earth Tech. 315 p.
SCIRE, J., STRIMAITIS, D.G., YAMARTINO, R.J. and
XIAOMONG, Z. 2000. A User’s Guide for CALPUFF
Dispersion Model (Version 5). Concord, MA: Earth
Tech, Inc. 512 p.
SESTAK, M., O'NEILL, S., FERGUSON, S., CHING, J. and
FOX, D.G. 2002. Integration of Wildfire Emissions
into Models-3/Cmaq with the Prototypes:
Community Smoke Emissions Modeling System
(CSEM) and BLUESKY. 2002 Models-3 User’s
Workshop, October 21-23, EPA, Research Triangle
Park, NC, USA.
SESTAK, M.L. and RIEBAU, A.R. 1988. SASEM - Simple
Approach Smoke Estimation Model. Tech. Note
382. U.S. Department of the Interior, Bureau of Land
Management. Denver, CO, USA. 31 p.
29
EUFIRELAB

SESTAK, M.L., MARLATT, W.E. and RIEBAU, A.R. 1988. WARD, D., ROTHERMEL, R. and BUSHEY, C. 1993.
VALBOX: Ventilated Valley Box Model. Unpublished Particulate matter and trace gas emissions from the
draft. U.S. Bureau of Land Management. 32 p. Canyon Creek Fire of 1988. In: Society of American
Foresters (Eds.), Proceedings of the 12th Fire and
SLOANE, C.S., WATSON, J., CHOW, J., PRITCHETT, L. and
Forest Meteorology, Georgia, United States of
RICHARDS L.W. 1991. Size-Segregated Fine Particle
America, pp. 62-76.
Measurements by Chemical Species and their
Impact on Visibility Impairment in Denver. WARD, D.E. 1999. Smoke From Wildland Fires. In:
Atmospheric Environment., 25A(5-6):1013-1024. Health Guidelines for Vegetation Fire Events -
Background Papers. Kee-Tai-Goh, Schwela D.,
TAPPER, N.J. and HESS, G.D. 1999. Forest Fire
Goldammer J.G. and Simpson O. (Eds). United
Emissions Dispersion Modelling for Emergency
Nations Environment Programme, Nairobi, World
Response Planning: Determination of Critical Model
Health Organization, Geneva, World Meteorological
Inputs and Processes. In: Health Guidelines for
Organization, Geneva, Institute of Environmental
Vegetation Fire Events - Background Papers. Kee-
Epidemiology, WHO Collaborating Centre for
Tai-Goh, Schwela D., Goldammer J.G. and Simpson
Environmental Epidemiology, Ministry of the
O. (Eds). United Nations Environment Programme,
Environment, Singapore. pp 70-85.
Nairobi, World Health Organization, Geneva, World
Meteorological Organization, Geneva, Institute of WARD, D.E. and HARDY, C. 1989. Organic and
Environmental Epidemiology, WHO Collaborating Elemental Profiles for Smoke from Prescribed Fires.
Centre for Environmental Epidemiology, Ministry of In: Watson J.G. (Ed.). Receptor Models in Air
the Environment, Singapore. pp 122-146. Resources Management: transactions of an
international specialty conference of the Air and
TRIJONIS, J.C., MALM, W.C., PITCHFORD, M., WHITE,
Waste Management Association, San Francisco,
W.H., CHARLSON, R. and HUSAR, R. 1991. Visibility:
CA. Pittsburg, PA, USA. pp. 299-321.
Existing and Historical Conditions - Causes and
Effects. In: Irving, P.M. (Ed.). Acidic Deposition: WHO/UNEP/WMO. 1999. Health Guidelines for
State of Science and Technology, Volume III: Vegetation Fire Events - Guideline Document.
Terrestrial, Materials, Health and Visibility Effects. SCHWELA D., GOLDAMMER J.G., MORAWSKA L.,
The U.S. National Acid Precipitation Assessment SIMPSON O. (Eds.). United Nations Environment
Program (State of science and technology report no. Programme, Nairobi, World Health Organization,
24). Washington, DC, USA. Geneva, World Meteorological Organization,
Geneva, Institute of Environmental Epidemiology,
USEPA. 1992. User's Manual for the Plume Visibility
WHO Collaborating Centre for Environmental
Model (PLUVUE II) (Revised). U.S. Environmental
Epidemiology, Ministry of the Environment,
Protection Agency, EPA-454/B-92-008.
Singapore.
USEPA. 1993. Interagency Workgroup on Air Quality
Modeling (IWAQM) Phase 1 Report: Interim
Recommendation for Modeling Long Range url 1: http://www.fs.fed.us/r6/fire/fastracs/
Transport and Impacts on Regional Visibility. U.S.
url 2: http://www.fs.fed.us/pnw/fera/vent/
Environmental Protection Agency, EPA-454/R-93-
015. url 3: http://www.fs.fed.us/fcamms/
USEPA. 1995. Interim Findings on the Status of url 4: http://www.arl.noaa.gov/smoke/index.html
Visibility Research. Office of Research and
Development, U.S. Environmental Protection url 5: http://www.fs.fed.us/land/wfas/
Agency, February.
USEPA. 1999. Science Algorithms of the EPA Models-
3 Community Multiscale Air Quality (CMAQ)
Modeling System. U.S. Environmental Protection
Agency, EPA/600/R-99/030.
VALENTE, J., MIRANDA, A.I., LOPES, A., BORREGO, C.
and VIEGAS, D.X. 2007. A local-scale modelling
system to simulate smoke dispersion. International
Journal of Wildland Fire. Accepted for publication.
WADE, D.D. and WARD, D.E. 1973. An Analysis of the
Air Force Bomb Range Fire. Research Paper SE-
105, U.S. Department of Agriculture, Forest Service,
Southeastern Forest Experiment Station. Ashville,
NC, USA. 38 p.

D-03-09.doc 30
EUFIRELAB

3.6 FIGURES

Figure 1 – CO concentration (µg.m-3) and wind fields obtained for the burning of a experimental fire
during GESTOSA 2004 experiments.

Figure 2 – Fire Consortia for the Advanced Modelling of Meteorology and Smoke (FCAMMS).
Available online in url 3.

D-03-09.doc 31
EUFIRELAB

Figure 3 – HYSPLIT archived smoke forecasts for South Central U.S. fires. Available in url 4.

Figure 4 – Example of a fire danger map obtained from the Wildland Fire Assessment System (WFAS).
Available online in url 5.

D-03-09.doc 32
EUFIRELAB
4 FIRE BEHAVIOUR PREDICTION
4.1 INTRODUCTION
Before stepping to the more specific subject of this
review, it is perhaps worthwhile to mention the broad
synthesis of André et al. (1992), covering virtually all
aspects of research on the physics of forest fires,
namely, mattering for /1.1-2/, /1.3/, /2.1-5/ and /3.1-3/.
In this review, four main criteria are used to
organise the state-of-the-art of research, based,
respectively, on:
- the concept of fire development phase; the concept
of fire regime 1;
- the main aspect of the fire behaviour that is
predicted; and
- the more or less applied character of the research.
Although they are used essentially in parallel, there
is some cross linking among these criteria.
4.2 FIRE DEVELOPMENT PHASE
Following ANDRÉ (1996), we can distinguish eight
main potential development phases in the history of a
general forest fire, named after:
- (1) ignition;
- (2) build-up;
- (3) full development of a specific fire regime;
- (4) (eventual) transition of fire regime;
- (5) decay;
- (6) extinguishing of the flame (for a flaming fire
front); (
- (7) smouldering combustion, till its extinction; and
- (8) cooling of the combustion residues to ambient
temperature.
The study of ignition (1st phase) is relevant for /1.3-
5/.
As it is a rather self-contained research area,
bearing few relations with fire spread studies; we do not
review it here (see, for instance, André et al. 1992).
The exceptions apply to:
- the influence of ignition conditions on the build-up
phase of the fire history (see below);
- the concept of temperature of ignition, which is used
by most semi-empirical fire spread models; and,
- in a less extent, the concept of ignition time delay,
which is seldom used (for instance, in the semi-
empirical fire spread model of Fons 1946).
During the build-up phase (2nd phase), the fire
progressively loses the memory, so to say, of the
ignition conditions, before attaining a fully developed
steady regime.
The duration of this phase is a matter of interest for
initial dispatch operations /1.2/ and for performing fire
experiments.
1
Terms in italic are defined in the DELFI Vocabulary.
D-03-09.doc
From a theoretical point of view, three effects that
are responsible for the time variation of the rate of
spread of the fire during this phase, have been studied,
although not very deeply:
- the establishing of the fire front depth (Emmons
1964);
- the establishing of a steady temperature field in the
part of the fuel bed (or fuel complex) still non-
burned, ahead of the fire line, in a reference frame
linked to the latter (Fujii et al. 1980); and the effect
of the fire line curvature characteristic of a point-
ignited fire (Cekirge 1978, Weber 1989).
Relevant experimental studies on this subject are:
MC ARTHUR (1966), JOHANSEN (1987) and MC ALPINE
(1988).
After the build-up phase, the fire becomes fully
developed (3rd phase), being characterised by a
(quasi-)steady behaviour which is exclusively
determined by the fuel bed and ambient conditions.
This is, by far, the most studied phase of forest fires.
Indeed, throughout the remaining sections of this
review, when not mentioned otherwise, it should be
understood that the works cited deal with the behaviour
of fire during this phase of its history.
It is also the richest phase of the fire history in what
concerns the diversity of possible behaviours of the fire.
In the latter regard, the concept of fire regime
introduced in the following Section (Section 3), can be
envisaged as a very convenient way of subdividing the
study of the fire in this phase.
Some fires pass through a fourth phase, which
consists in, a transition of regime caused by a
significant change of the fuel bed or ambient
conditions.
This phase is usually rather brief due to strong non-
linear effects that come into play.
Its complexity explains the relatively few research
done on it.
Below, two particularly important fire regime
transitions are mentioned: flare up from a smouldering
fire, and blow-up of a large crown or bush fire from a
less intense surface fire.
When the conditions of the fuel bed (i.e. high fuel
moisture or very low fuel bed porosity) or the ambient
(i. e. strong opposing wind) turn it difficult for the fire to
spread, the fire may enter into an unstable decay
phase (5th phase).
The rate of spread shows a very high sensitivity to
the fuel bed or ambient pertinent parameter.
Although the fire is non-steady in this phase of its
history, for brief periods of time, it behaves in a way
that is similar to a steady fire regime that belongs to a
class of, so called, marginal fire regimes, the research
of which is reviewed in Section 3.
33
EUFIRELAB
The decay phase may ultimately lead to the flame
extinction (6th phase).
The study of the fire in this phase is especially
interesting associated with the use of fire suppression
means (see /1.2/ and /1.7/, and also ANDRÉ et al.
1992).
Usually, the flame extinction phase is followed by a
seventh phase, characterised by a smouldering
combustion regime in the fire front, which
- proceeds at the surface of the solid fuel bed
particles (usually, during this phase, already
practically reduced to char), and
- requires much less oxygen (air) and
- has a much lower burning rate, and, so, a much
higher residence time, than flaming combustion (cf.
CHANDLER et al. 1983).
The study of this phase of the fire is relevant for
(see /1.2/, /1.5/, /2.3/, /3.1/):
- the soil heating (p.e. DIMITRAKOUPOULOS and
MARTIN 1990);
- the smoke generation (p.e. KANURY 1976, RASBASH
and DRYSDALE 1982); and
- the rekindling of flame (flare up) due to sudden
changes of the fuel bed or ambient conditions (FONS
1950, Chandler et al. 1983).
The final cooling of the combustion residues (8th
phase) has not attracted much research attention (see,
however, /3.1/).
D-03-09.doc
4.3 FIRE REGIME
The concept of fire regime is thoroughly defined and
explored in ANDRÉ (1996).
As it is said above, this concept allows a finer
discrimination of the fire behaviour during its fully
developed phase.
Although it can also be used, for short periods of
time, during other, non-steady phases of the fire
history, such as the build-up and the decaying phases.
Below, four main criteria of subdivision of the fire
regimes, not entirely independent, are used to review
the research.
These criteria are based, respectively:
- on: the combustion regime in the fire front, which
can be glowing or smouldering, with laminar or
turbulent diffusion flames, or even, although seldom,
with a premix flame (cf. ROTHERMEL 1991, on tree
torching);
- the type of forest fuels that carry the front, which
can give place to ground fires, surface fires and
crown fires that are either free or dependent on
surface fires spreading underneath;
- the sensitivity of the fire response to a change in a
given input parameter of the fuel bed or of the
ambient;
- the order of magnitude of the values of certain
properties of the fire line, such as, the rate of spread
(slow and fast fire regimes) or the fire line intensity
(low, medium and high intensity fire regimes).
The class of, so called, marginal fire regimes
(ANDRÉ 1998b) deserves a special reference, for its
interest in the study of:
- the transition between the ignition and build-up
phases, and
- the decaying phase of the fire history.
An interesting open question concerning forest fire
regimes is the (eventual) existence of multiple fire
regimes for the same conditions of the fuel bed and
ambient, depending on the history of fire prior to the
fully developed phase and, in particular, depending on
its ignition conditions (EMMONS and SHEN 1971,
STEWARD 1974, PALMER and NORTHCUTT 1975, THOMAS
1967,1971).
34
EUFIRELAB
4.4 ASPECTS OF THE FIRE BEHAVIOUR
PREDICTED
4.4.1 Main fire front
4.4.1.1 Ground fires
Ground fires, perhaps because they are the least
intense of all the forest fire regimes (due to the
smouldering combustion regime that characterises it,
differently from the much more energetic flaming
combustion regimes that characterise the surface and
crown fires), have not attracted much research
attention.
On this subject, besides the references given above
about the smouldering combustion phase of the history
of a more general fire (7th phase), see: Mc MAHON et al.
(1980), WILLIAMS (1982) and WADE (1984).
4.4.1.2 Surface fires
This is the area of forest fire physics research in
which more theoretical and experimental work has
been done.
A useful reference covering practically all types of
fire spread models mostly dealing with the main fire
front behaviour of surface forest fires, is the recent
review of ANDRÉ and VIEGAS (1999), which is
addressed to non-specialists.
A major division that ought to be done in the
research in this area respects the distinction between
the modelling of a small section of the fire line during
small periods of time, and the modelling of the whole
fire line for long time periods.
The former type of predictions are said to have a
local character while the latter have a global character.
4.4.2 Behaviour of a small section of the fire
line (local prediction)
Almost all studies included in this first group apply to
a 2D (i.e., straight line and infinite), fully developed and
flaming fire front, spreading through a statistically
homogeneous fuel bed, on plane but inclined terrain,
under a uniform wind field.
Moreover, all the properties of the fuel bed and the
ambient are constant and, if they exist, the directions of
maximum slope and of wind velocity are both
perpendicular to the straight fire line.
Closest to the empirical bottom of the modelling
scale, we have the fire behaviour models based on
wildfire observations and field tests, typically with a
scope of application rather restrict, such as:
- in Australia, the works reported in LUKE and MC
ARTHUR (1978), GILL and NOBLE (1989) and
CATCHPOLE (1998);
- in Canada, large parts of the work underlying the
Canadian Forest Fire Behaviour Prediction System
(STOCKS et al. 1989); and,
- in the USA, the early work of CURRY and FONS
(1940).
D-03-09.doc
Upper, in a second level of the modelling scale,
there are the empirical models based on laboratory
experiments.
However, in the first place, it should be stressed that
most of the laboratory work has other goals, such as:
- studying the phenomena that control the fire spread
in certain fire regimes;
- investigating the influence of given input parameters
of the fuel bed and ambient (such as: the thickness
and moisture of the fuel bed particles, the fuel load
and the fuel bed porosity, the wind speed or the
slope), either in an isolated fashion or in small
groups, on given output parameters characterising
the fire front behaviour (such as: the rate of spread,
the depth and residence time of the fire front, and
the flame length and inclination),
- sometimes leading to the proposal of particular
empirical correlation; or validating any fire behaviour
model of the semi-empirical (physical incomplete) or
comprehensive (physical complete) type.
STEWARD (1974), ANDRE et al. (1992), VAZ (1997)
and CATCHPOLE and CATCHPOLE (1998) refer numerous
studies of this kind.
NELSON and ADKINS (1988) and CARRIER et al.
(1991) deal specifically with Dimensional Analysis and
Similitude
Theory problems posed by these type of
experiments.
One of the few, and, clearly, the most outstanding
laboratory empirical fire spread model is the model of
ROTHERMEL (1972), which is the core of the well known
software BEHAVE (ANDREWS 1986, 1989).
Going up the modelling scale, we find, at a third
level, the physical-incomplete or semi-empirical fire
spread models.
These models are back-up on a set of physical
laws, the most important of which expresses the
conservation of energy in a control volume located in
the part of the fuel bed not yet burned, ahead of the fire
line
The rate of spread of the fire, which is the main
output of the models, appears as an integral parameter
(albeit, eventually hidden) in this equation.
The equation explicitly includes two sets of terms
associated, respectively, with heat sink and heat flux
phenomena, but the positive heat flux terms,
representing an income of energy to the control
volume, depend on a third set of heat source
phenomena occurring at the fire front.
In fact, even ROTHERMEL model is based on such a
law, written in the simplest possible (space-time
integral) form, but, contrary to the semi-empirical
models, does not model in a physical way the heat flux
terms.
However, the semi-empirical models cannot be
considered as physical-complete models because they
do not model the heat source terms.
35
EUFIRELAB
The following are review works specifically dealing
with this class of fire spread models: STEWARD (1974),
CATCHPOLE and DE MESTRE (1986), WEBER (1991) and
CATCHPOLE (1994).
ANDRE et al. (1992) contains also interesting
material on this regard.
A first group of models of this type, treating the fire
spread in a space-time continuous way, is: EMMONS
(1964), THOMAS et al. (1964), HOTTEL et al. (1965),
FANG (1966), VAN WAGNER (1967), ALBINI (1967),
THOMAS (1967), FANG (1969), ANDERSON (1969),
BERLAD (1970), SANDHU (1970), FRANDSEN (1971),
THOMAS (1971), STEWARD (1971), HOTTEL et al. (1971),
PAGNI and PETERSON (1973), TELISIN (1974), PAGNI
(1975), CEKIRGE (1978), FUJII et al. (1980), ALBINI
(1985,1986), DE MESTRE et al. (1989) and WEBER
(1989).
There is a second group of models in which the fire
spread is modelled in a discontinuous way, as a
succession of jumps from particle to particle along the
fuel bed.
Although these models seem more adapted to cope
with very low porosity fuel beds, in any case, they can
always be converted into formally continuous fire
spread models.
Here is a list of works in this group: FONS (1946),
VOGEL and WILLIAMS (1970), EMMONS and SHEN (1971),
STEWARD and WAIBEL (1973) and WEBER (1990).
The software FIRELAB (GUARNIERI et al. 1998),
under development, implements some of the former
models in an easy-to-use way.
Finally, in a fourth level of the modelling scale, there
is still another group of fire spread models that can be
already classified as physical complete models (see the
DELFI review of LARINI).
It is a striking fact that the rate of production of new
semi-empirical models practically comes to zero in the
nineties, while the one of physical complete models
begins to rise.
4.4.3 Behaviour of the whole fire line (global
prediction)
Let us now step to the global scope of prediction of
the fire behaviour, in which the whole fire line is
followed for long periods of time, paying attention to the
significant space and time changes of the fuel bed and
ambient conditions along the fire line, during the time
interval of prediction.
In this area, not accounting for the models of the
physical complete type (see the DELFI review of
LARINI), practically all the research can be cast into the
theoretical framework of ANDRÉ and VIEGAS (1998b),
which is summarised in ANDRÉ and VIEGAS (1999) (see
also the critical proposal of VIEGAS et al. 1998).
D-03-09.doc
However, there also exist some works with a deep
statistical nature, using concepts and tools of Chaos
Theory, such as:
- percolation models,
- cellular automata models and
- fractal geometry.
They have been applied to study the behaviour of
forest fires at a global level, with the purpose, for
instance:
- of identifying critical behaviours (eventually
associated with marginal fire regimes) or
- of determining the fractal dimension of the burned
area.
These works, essentially distinct from the works of
the main stream, are not reviewed here (see the early
work of VON NIESSEN and BLUMEN 1988, and the recent
review of DUARTE 1997).
4.4.4 Large crown fires and bushfires
Large crown fires and bushfires are much less
studied than lower intensity surface fires.
See ANDRE et al. (1992) for a more extensive review
of this subject.
The initiation of a fire regime of this class is usually
associated with a regime transition from lower intensity
surface fires, named blow-up.
It is studied in the following works: BYRAM (1954),
SHAEFER (1957), MOLCHANOV (1957), VAN WAGNER
(1964, 1977), EMMONS (1966), BROWN and DAVIS
(1973), DIETERICH (1976), BROTAK (1978), LUKE and MC
ARTHUR (1978), WILLIAMS (1982), CARRIER et al. (1985),
HAINES (1988), SIMARD and EENIGENBURG (1990).
Tentative identifications of different fire regimes
within this class are done by: BYRAM (1966), LEE
(1972), PALMER and NORTHCUTT (1975) and
ROTHERMEL (1991).
Dimensional Analysis and Similitude Theory
considerations relevant to the modelling of large forest
fires can be found in BYRAM (1966) and WILLIAMS
(1969).
Some authors studied the strong flows induced by
this type of fires, such as, the main fire plume and
different vortex flow structures, but as this subject
pertains more to /2.4/, we do not review it here.
Regarding the fire spread models that have been
proposed specifically to model large forest fires, we
have:
- the empirical models of VAN WAGNER (1989) and
ROTHERMEL (1991), which are incorporated,
respectively, in the Canadian Forest Fire Behaviour
Prediction System and in the software package
BEHAVE (see Section 5);
- the semi-empirical model of ALBINI (ALBINI and
STOCKS 1986); and, finally,
- some models of the physical complete type (see the
DELFI review of LARINI).
An on-going important field experimental program
on crown fires is reported in ALEXANDER et al. (1998).
36
EUFIRELAB
4.4.5 Spotting
Spotting, albeit being present in any type of flaming
forest fire, even in the least intense ones (cf. STEWARD
1974), only in medium-to-high intensity fires becomes a
really important phenomenon in the context of /1.2/ and
/2.5/, which ought to be predicted in addition to the
behaviour of the main fire front (see Section 4.1).
However, this is a relatively self-contained research
problem seldom addressed.
Following ANDRÉ et al. (1992), we can distinguish
three main sub-problems in the study of spotting:
- the production of firebrands in the main fire front (or,
more generally, in any spotting source, such as can
be an isolated torching tree);
- the transport of the firebrands by the flow of hot
gases and wind, from the fire line towards the still
non-burned part of the fuel bed, beginning with the
lifting of a firebrand in the fire plume and ending with
its gravity fall across the wind flow; and
- the (eventual) ignition of a secondary fire spot,
ahead of the main front.
The study of the transport problem, obviously,
benefits from independent modelling both of the fire
plume (see, p. e., LEE and EMMONS 1961, MORTON
1965, NIELSON and TAO 1965, SMITH 1967, BYRAM and
NELSON 1974, WILLIAMS 1982), and of the wind flow
over complex terrain /2.4/.
Similarly, the investigation of secondary fire spots
can take profit of the research on the more general
subject of ignition of forest fires (see Section 2).
Specifically addressing the spotting issue, we have:
the wildfire observations reported by LUKE and MC
ARTHUR (1978); the experimental studies of TARIFA et
al. (1965) and STEPHEN and WRIGHT (1974); and the
semi-empirical models of ALBINI (1979, 1982).
4.4.6 Fire whirls
Fire whirls are large scale coherent vortex flow
structures – according to EMMONS (1964).
These vortexes can be assimilated to cylinders with
(diameter x height):
- ranging from (10 cm x 30 cm), for a small fire whirl,
- to (100 m x 300 m), for a strong fire tornado 2
- carrying inside flames and firebrands, that
sometimes detach from the fire front and wander
around for a while till they dissipate.
As this seems to be a rather rare and fortuitous
phenomenon in most forest fires, besides seldom
posing fire safety problems 3, we do not review its
research here (such a review can be found in ANDRÉ et
al. 1992).
2
Probably, forest fires can never become intense enough
to generate such a fire tornado, which, on the other hand,
has been observed in large industrial and urban fires.
3
Personal communication done to the author by Richard
st
ROTHERMEL, at the 1 International Conference on Forest
Fire Research (19-22 November 1990, Coimbra,
Portugal).
D-03-09.doc
4.5 APPLIED RESEARCH PRODUCTS
To simulate with a particular model the behaviour of
a fire front in natural dynamic scenery of fuel bed and
ambient conditions, either on virtual, historical, real-
time or prognostic grounds, one must solve two main
problems:
- to get the input information required by the model,
and
- to perform all the due mathematical computations.
In accordance, to facilitate these tasks for the
operational user, an applied research effort has been
done by the forest services of some countries
(particularly the USA), by enterprises and by research
groups, ultimately leading to the construction of more or
less elaborated and user-friendly software packages.
Some of these software systems are in the public
domain, others are commercial products, and still
others have an indefinite statute.
Many of these software systems, in addition to a
module specifically dealing with the prediction of the
fire behaviour, which incorporates some of the models
reviewed above, contain other modules that help the
user:
- in organising activities of fire management /1.1/,
- fire suppression /1.2/,
- fire detection /1.3/,
- fire risk /1.4/, or
- even other more or less specialised fire activities
(/1.5/, /3.6/);
- in the fuel characterisation /2.1/ or
- in wind modelling /2.4/; or
- in the managing of smoke (/2.3/, /3.1/).
As all these functions pertain to other topics of
DELFI Index, we concentrate here on systems
possessing an important or original fire behaviour
module.
The most important systems that give only local fire
behaviour predictions are:
- CFBPS Canadian Forest Fire Behaviour Prediction
System and
- the software package BEHAVE.
Regarding the numerous Australian subsystems see
(CATCHPOLE 1998).
Although aimed at laboratory simulations more than
at operational applications, the software FIRELAB
should also be mentioned.
All the existing software systems that produce fire
behaviour predictions with a global character and
incorporate empirical or semi-empirical fire spread
models can ultimately be framed within the theory (see
Section 4.1.2).
In fact, a rather general software kernel of
implementation of this theory is presently under
development.
Moreover, practically all systems use BEHAVE as
Local Fire Spread Model.
37
EUFIRELAB
Here is an illustrative list of such systems:
- FMISTM, FIRESTATIONTM, GEOFOGO and
CARDIN, using slightly different variants of the
algorithm of DIJKSTRA to implement a Global Fireline
Propagation Model based on the principle of the
fastest path of propagation; and
- FARSITE, which implements HUYGENS' principle as
the basis of a Global Fireline Propagation Model.
Some attempts were made to adapt it to European
shrublands (ARCA et al. 2006, see also TECNOMA-FSE
in deliverable D-03-06).
These systems are in constant evolution to take
better into account external and vegetation conditions
(BUTLER et al. 2006).
Concerning detailed physical modelling, the WFDS
simulator is in under development, as stated in section
2.4.
(Review prepared under an Associated Contract to
the implementation of the CONCERTED ACTION
DELFI)
4.6 FUTURE PROSPECTS
To simulate with a particular model the behaviour of
a fire front in a natural dynamic scenery of fuel bed and
ambient conditions, either on virtual, historical, real-
time or prognostic grounds, one must solve two main
problems:
- to get the input information required by the model,
and
- to perform all the due mathematical computations.
As the authors stated in section 2.4, physical and
semi-empirical approaches are complementary ones
and they should be mixed to describe such a situation
(at this point of our knowledge on the fire behaviour).
D-03-09.doc
4.7 REFERENCES
ALBINI, F. A. 1967. A physical model for fire spread in
brush. Proc. 11th International Symposium on
Combustion, pp. 553-560.
ALBINI, F. A. 1976a. Combining wind and slope effects
on spread rate. United States Department of
Agriculture, Forest Service, Northern Forest Fire
Laboratory (Missoula). Memorandum to R. C.
ROTHERMEL (January 19).
ALBINI, F. A. 1976b. Computer-based models of fire
behaviour: a user's manual. United States
Department of Agriculture, Forest Service,
Intermountain Forest and Range Experiment Station
(Ogden, Utah). 68 pp..
ALBINI, F. A. 1979. Spot fire distance from burning trees
– A predictive model. United States Department of
Agriculture, Forest Service, Intermountain Forest
and Range Experiment Station (Ogden, Utah).
General Technical Report INT-56.
ALBINI, F. A. and R. G. BAUGHMAN. 1979. Estimating
windspeeds for predicting wildland fire behaviour.
United States Department of Agriculture, Forest
Service (Ogden, Utah). Research Paper INT-221.
ALBINI, F. A. 1982. Response of free burning fires to
non-steady wind. Combustion Science and
Technology 29: 225-241.
ALBINI, F. A., LATHAN, D. J. and R. G. BAUGHMAN. 1982.
Estimating up-slope convective windspeeds for
predicting wildland fire behaviour. United States
Department of Agriculture, Forest Service (Ogden,
Utah). Research Paper INT-257.
ALBINI, F. A. 1985. A model for fire spread in wildland
fuels by radiation. Combustion Science and
Technology 42:229-258.
ALBINI, F. A. 1986. Wildland fire spread by radiation – A
model including fuel cooling by natural convection.
Combustion Science and Technology 45:101-113.
ALBINI, F. A. and B. J. STOKES. 1986. Predicted and
observed rates of spread of crown fires in immature
jack pine. Combustion Science and Technology
48:65-76.
ALBINI, F. A. 1996. Iterative solution of the radiation
transport equations governing spread of fire in
wildland fuel. Physika Goreniya I Vzryva 32(5): 71-
82.
ALEXANDER, M. E., STOCKS, B. J., WOTTON, B. M. and
R. A. LANOVILLE. 1998. Proc. III International
Conference on Forest Fire Research / 14th
Conference on Fire and Forest Meteorology (Luso,
16/20 November 1998), Vol. I, pp. 83-112.
ANDERSON, D., CATCHPOLE, E., DE MESTRE, N. and T.
PARKES. 1982. Modelling the spread of grass fires.
Journal of the Australian Mathematical Society,
Series B, 23:451-466.
38
EUFIRELAB
ANDERSON, H. E. 1968. Sundance fire: an analysis of
fire phenomena. United States Department of
Agriculture, Forest Service, Intermountain Forest
and Range Experiment Station (Ogden, Utah).
General Technical Report INT-122. 39 pp..
ANDERSON, H. E. 1969. Heat transfer and fire spread.
United States Department of Agriculture, Forest
Service. Research Paper INT-69. 20 pp..
ANDERSON, H. E. 1982. Aids to determining fuel models
for estimating fire behaviour. United States
Department of Agriculture, Forest Service. General
Technical Report INT-122.
ANDERSON, H. E. 1983. Predicting wind-driven wildland
fire size and shape. United States Department of
Agriculture, Forest Service (Ogden, Utah).
Research Paper INT-305. 26 pp..
ANDRÉ, J. C. S., LOPES, A. G. and D. X. VIEGAS. 1992.
A broad synthesis of research on physical aspects
of forest fires. Fluid Mechanics Group, Mechanical
Engineering Department, Faculty of Science and
Technology, University of Coimbra. Cadernos
Científicos sobre Incêndios Florestais 3. 148 pp..
ANDRE, J. C. S. 1996. A theory on the propagation of
surface forest fire fronts. PhD. Thesis (in
portuguese). Mechanical Engineering Department,
Faculty of Science and Technology, University of
Coimbra. 330 pp..
ANDRÉ, J. C. S. and D. X. VIEGAS. 1998a. Modelling of
the propagation of a forest fire front: appraisal of an
algorithm in common use (DIJKSTRA'S algorithm).
Proc. 2nd Numerical Meeting on Partial Differential
Equations (25-27 Feb, Coimbra, Portugal).
ANDRÉ, J. C. S. and D. X. VIEGAS 1998b. An unifying
theory on the propagation of the fire front of surface
forest fires. Proc. III International Conference on
Forest Fire Research / 14th Conference on Fire and
Forest Meteorology (Luso, 16/20 November 1998),
Vol. I, pp. 259-279.
ANDRÉ, J. C. S. and D. X. VIEGAS 1999. Forest fires
spread models: state-of-the-art for non-specialists.
(In portuguese. Draft version.)
ANDREWS, P. L. 1986. BEHAVE: fire behaviour
prediction and fuel modelling system – – BURN
subsystem, Part 1. United States Department of
Agriculture, Forest Service (Ogden, Utah). General
Technical Report INT-194. 130 pp..
ANDREWS, P. L. and G. A. MORRIS. 1987. Equations for
wildland fire behaviour prediction: a summary.
United States Department of Agriculture, Forest
Service, Intermountain Fire Sciences Laboratory
(Missoula, Montana).
ANDREWS, P. L. and C. H. CHASE. 1989. BEHAVE: fire
behaviour prediction and fuel modelling system -
BURN subsystem, Part 2. United States Department
of Agriculture, Forest Service (Ogden, Utah).
General Technical Report INT-260. 93 pp..
D-03-09.doc
ANDREWS, P. L. and L. S. BRADSHAW. 1990.
RXWINDOW: defining windows of acceptable
burning conditions based on desired fire behaviour.
United States Department of Agriculture, Forest
Service, Intermountain Research Station (Ogden,
Utah). General Technical Report INT-273. 54 pp..
ANDREWS, P. L. and C. D. BEVINS. 1998. Update and
expansion of the BEHAVE fire behaviour prediction
system. Proc. III International Conference on Forest
Fire Research / 14th Conference on Fire and Forest
Meteorology (Luso, 16/20 November 1998), Vol. I,
pp. 733-740.
ARCA B., DUCE P., PELLIZZARO G., LACONI M., SALIS M.,
SPANO D. 2006. Evaluation of FARSITE simulator in
Mediterranean shrubland. In Proc. V Int. Conf. on
Forest Fire Research, VIEGAS, D.X. (Ed.), Figueira
da Foz, 27-30 Nov. Elsevier Publishers.
BARROS, F. J. and G. L. BALL. 1998. Fire modelling
using dynamic structure cellular automata. Proc. III
International Conference on Forest Fire Research /
14th Conference on Fire and Forest Meteorology
(Luso, 16/20 November 1998), Vol. I, pp. 879-888.
BERLAD, A. L. 1970. Fire spread in solid fuel arrays.
Combustion and Flame 14:123-136.
BRADSHAW, L. S., DEEMING, J. E., BURGAN, R. E. and J.
D. COHEN. 1984. The 1978-National Fire Danger
Rating System: Technical documentation. United
States Department of Agriculture, Forest Service.
General Technical Report INT-169. 44 pp..
BROTAK, E. A. 1978. The Bass-River fire – Weather
conditions associated with a killer fire. Proc. 5th
Symposium on Forest Meteorology (Ottawa,
Ontario, Canada).
BROWN, A. A. and K. P. DAVIS. 1973. Forest fire: control
and use (2nd Ed.). Ed. McGraw-Hill.
BRUCE, H. D., PONG, W. Y. and W. L. FONS. 1961. The
effect of density and thermal diffusivity of wood on
the rate of burning of wood cribs. United States
Department of Agriculture, Forest Service.
Technical Paper N. 63.
BURGAN, R. E. and R. C. ROTHERMEL. 1984. BEHAVE:
fire behaviour prediction and fuel modelling system -
FUEL subsystem. United States Department of
Agriculture, Forest Service (Ogden, Utah). General
Technical Report INT-167. 126 pp.
BUTLER B., FORTHOFER J., FINNEY M., MCHUGH C.,
STRATTON R., BRADSHAW L. 2006. The impact of
high resolution wind field simulations on the
accuracy of fire growth predictions. In Proc. V Int.
Conf. on Forest Fire Research, VIEGAS, D.X. (Ed.),
Figueira da Foz, 27-30 Nov. Elsevier Publishers.
BYRAM, G. M. 1954. Atmospheric conditions related to
blow-up fires. United States Department of
Agriculture, Forest Service, Southeast Forest and
Range Station. Paper SE-35.
39
EUFIRELAB

BYRAM, G. M. 1959. In: DAVIS, K. P. 1959. Forest fire: CATCHPOLE, W. A., CATCHPOLE, E. A., ROTHERMEL, R.
control and use. (1st Ed.) Ed. McGraw-Hill. C., MORRIS, G. A., BUTLER, B. W. and D. J. LATHAM.
Chapters: Combustion of forest fuels and Forest fire 1998. Rate of spread of free-burning fires in woody
behaviour. fuels in a wind tunnel. Combustion Science and
Technology 131: 1-37.
BYRAM, G. M., CLEMENTS, H. B., ELLIOT, E. R. and P. M.
GEORGE. 1964. An experimental study of model CEKIRGE, H. M. 1978. Propagation of fire fronts in
fires. United States Department of Agriculture, forests. Computational and Mathematical
Forest Service, South-eastern Forest and Applications 4:325-332.
Experiment Station, Southern Forest Fire
CHANDLER, C., CHENEY, P., THOMAS, P., TRABAUD, L.
Laboratory (Macon, Georgia). Technical Report N.
and D. WILLIAMS. 1983. Fire in forestry - Forest fire
3.
behaviour and effects. Ed. John Wiley & Sons. Vol.
BYRAM, G. M. 1966. Scaling laws for modelling mass 1.
fires. Pyrodynamics 4: 271-284.
CHENEY, N. P., GOULD, J. S. and W. R. CATCHPOLE.
BYRAM, G. M., CLEMENTS, H. B., BISHOP, M. E. and R. 1998. Prediction of fire spread in grasslands.
M. NELSON. 1966. An experimental study of model International Journal od Wildland Fire 8 (1): 1-13.
fires – Project Fire Model. United States Department
COUNTRYMAN, C. M. 1983. Ignition of grass fuels by
of Agriculture, Forest Service, South-eastern Forest
cigarettes. Fire Management Notes 44: 3-7.
and Experiment Station, Southern Forest Fire
Laboratory (Macon, Georgia). Final Report. CURRY, J. R. and W. L. FONS. 1940. Forest fire
behaviour studies. Mechanical Engineering 62:219-
BYRAM, G. M. and R. M. NELSON. 1974. Buoyancy
225.
characteristics of a fire heat source. Fire
Technology 10(1): 68-79. DE MESTRE, N., CATCHPOLE, E., ANDERSON, D. and R.
ROTHERMEL. 1989. Uniform propagation of a planar
CALL, P. T. 1997. Run-time optimisation of a radiation
fire front without wind. Combustion Science and
driven crown fire model. MSc. Thesis. Montana
Technology 65:231-244.
State University (Bozeman, Montana). 86 pp..
DIETERICH, J. H. 1976. Jet stream influence on the
CALL, P. M. and F. A. ALBINI. 1997. Aerial and surface
Willow fire. Fire Management Notes 37: 6-8.
fuel consumption in crown fires. International
Journal of Wildland Fire 7: 259-264. DIMITRAKOPOULOS, A. P. and R. E. MARTIN. 1990.
Measuring and modelling soil temperatures during
CARRIER, G. F., FENDELL, F. E. and P. S. FELDMAN.
simulated wildland fire conditions. Proc.
1985. Firestorms. Journal of Heat Transfer 107: 19-
International Conference on Forest Fire Research
27.
(Coimbra). Communication B.21. 17 pp..
CARRIER, G. F., FENDELL, F. E. and M. F. WOLFF. 1991.
DUARTE, J. A. M. S. 1997. Bushfire models and their
Wind-aided fire spread across arrays of discrete fuel
phase transitions. International Journal of Modern
elements. I theory. Combustion Science and
Physics C (...)
Technology 75: 31-51.
EFTICHIDIS, G., MARGARITIS, E., SFIRIS, A. and V.
CATCHPOLE, E.A., DE MESTRE, N. J. and A. M. GILL.
VARELA. 1998. Fire management information
1982. Intensity of fire at its perimeter. Australian
systems: FMIS. Proc. III International Conference
Forestry Research 12:47-54.
on Forest Fire Research / 14th Conference on Fire
CATCHPOLE, E. A. and N. J. DE MESTRE. 1986. Physical and Forest Meteorology (Luso, 16/20 November
models for a spreading line fire. Australian Forestry 1998), Vol. II, pp. 2641-2642.
49(2):102-111.
EMMONS, H. W. 1964. Fire in the forest. Fire Research
CATCHPOLE, E. A. 1995. A review of physical fire Abstracts and Reviews 5(3):163--178.
models, 1945-1975. In: Elements of Bushfire
EMMONS, H. W. 1966. Fundamental problems of free
Science. Vol. 1: Fire Behaviour Models (PACKHAM,
burning fire. Fire Research Abstracts and Reviews
D. R., CATCHPOLE, E. A., CATCHPOLE, W. R.,
8(1):1-17.
KRUSEL, N. and N. VINEY, Eds.). 44 pp..
EMMONS, H. W. and T. SHEN. 1971. Fire spread in
CATCHPOLE, W. C. 1998. Fire behaviour research in
paper arrays. Proc. 13th International Symposium
Australia. Photocopies of the transparencies
on Combustion. pp. 917-926.
presented at the Short Course on Forest Fire
Behaviour Modelling (Luso, 21/22 November). 25 FANG, J. B. 1966. Fire spreading studies. MSc. Thesis.
pp.. Chemical Engineering Department, University of
New Brunswick (New Brunswick, Canada). 115 pp..
CATCHPOLE, W. C. and T. CATCHPOLE. 1998. Laboratory
based models. Photocopies of the transparencies
presented at the Short Course on Forest Fire
Behaviour Modelling (Luso, 21/22 November). 26
pp..

D-03-09.doc 40
EUFIRELAB

FANG, J. B. 1969. An investigation of the effect of FRANDSEN, W. H. and P. L. ANDREWS. 1979. Fire
controlled wind on the rate of fire spread. PhD. behaviour in nonuniform fuels. United States
Thesis. Chemical Engineering Department, Department of Agriculture, Forest Service (Ogden,
University of New Brunswick (New Brunswick, Utah). Research Paper INT-232.
Canada). 169 pp..
FUJII, N., HASEGAWA, J., PALLOP, L. and Y. SAKAWA.
FERNANDEZ-PELLO, A. C. and T. HIRANO. 1983. 1980. A non-stationary model of fire spreading.
Controlling mechanisms of flame spread. Applied Mathematics Modelling 4:176-180.
Combustion Science and Technology 32 (1-4): 1-32.
GILL, A. M. and I. R. NOBLE. 1989. Bibliography of fire
FEUNEKES, U. 1991. Error analysis in fire simulation ecology in Australia. Bushfire Council of New South
models. MSc. Thesis. Department of Forest Wales (2nd Ed.).
Research, University of New Brunswick (New
GREEN, D. G. 1983. Shapes of simulated fires in
Brunswick, Canada). 75 pp..
discrete fuels. Ecological Modelling 20: 21-32.
FINNEY, M. A. 1998. FARSITE: Fire Area Simulator –
GRISHIN, A. M., GRUZIN, A. D. and V. G. ZVEREV. 1983.
Model development and evaluation. United States
Mathematical modelling of the spreading of high-
Department of Agriculture, Forest Service, Rocky
level forest fires. Sovietic Physical Doklady 28:328-
Mountain Research Station. Research Paper
330.
RMRS-RP-4. 47 pp.
GRISHIN, A. M. 1984. Steady state propagation of the
FINNEY, M. A. and P. L. ANDREWS. 1998. Application
front of a high level forest fire. Sovietic Physical
and status of the FARSITE fire area simulator. Proc.
Doklady 29: 917-919.
III International Conference on Forest Fire Research
/ 14th Conference on Fire and Forest Meteorology GUARNIERI, F., ANDERSEN, C. K., OLAMPI, S. and N.
(Luso, 16/20 November 1998), Vol. I, pp. 755-760. CHAMBINAUD. 1998. FIRELAB: towards a problem
solving environment to support forest fire behaviour
FONS, W. L. 1946. Analysis of fire spread in light forest
modelling. Proc. III International Conference on
fuels. Journal of Agricultural Research 72:93-121.
Forest Fire Research / 14th Conference on Fire and
FONS, W. L. 1950. Heating and ignition of small wood Forest Meteorology (Luso, 16/20 November 1998),
cylinders. Industrial Engineering Chemistry Vol. I, pp. 483-496.
42(10):2130-2133.
HAINES, D. A. 1988. A lower atmosphere severity index
FONS, W. L., BRUCE, H. D., PONG, W. Y. and S. S. for wildland fire. Natural Weather Digest 13 (2): 23-
RICHARDS. 1960. Project Fire Model - An 27.
experimental study of model fires. United States
HOTTEL, H. C., WILLIAMS, G. C. and F. R. STEWARD.
Department of Agriculture, Forest Service, Pacific
1965. The modelling of fire spread through a fuel
Southwest Forest and Experiment Station (Macon,
bed. Proc. 10th International Symposium on
Georgia). Summary Progress Report.
Combustion, pp. 997-1007.
FONS, W. L., CLEMENTS, H., ELLIOT, E. R. and P. M.
HOTTEL, H. C., WILLIAMS, G. C. and G. K. KWENTUS.
GEORGE. 1962. Project Fire Model - An
1971. Fuel preheating in free-burning fires. Proc.
experimental study of model fires. United States
13th International Symposium on Combustion, pp.
Department of Agriculture, Forest Service, South-
963-970.
eastern Forest and Experiment Station (Macon,
Georgia). Summary Progress Report II. JOHANSEN, R. W. 1987. Ignition patterns and prescribed
fire behaviour in southern pine stands. Georgia
FONS, W. L., CLEMENTS, H. B. and P. M. GEORGE. 1963.
Forest Commission (USA), Research Division.
Scale effects on propagation rate of laboratory crib
Georgia Forest Research Paper.
fires. Proc. 9th International Symposium on
Combustion, pp. 860-866. JORBA, J., ANDRÉ, J. C. S., J., MARGALEF, T., LUQUE, E.
and D. X. VIEGAS. 1999. A framework for forest fire
Forestry Canada Fire Danger Group. 1992.
propagation simulation. Extended abstract
Development and structure of the Canadian Forest
submitted to the International Symposium on Forest
Fire Behaviour Prediction System. Forestry Canada
Fires: Needs & Innovations (18-19 November 1999,
(Ottawa, Ontario). Information Report ST-X-3. 63
Athens, Greece). (In preparation.)
pp..
KANURY, A. M. 1976. The science and engineering of
FRANDSEN, W. H. 1971. Fire spread through porous
hostile fires. Fire Research Abstract and Reviews
fuels from the conservation of energy. Combustion
18:72-96.
and Flame 16:9-16.
KONEV, E. V. and A. I. SUKHININ. 1977. The analysis of
FRANDSEN, W. H. 1973. Effective heating of fuel ahead
flame spread through forest fuel. Combustion and
of a spreading fire. United States Department of
Flame 28: 217-223.
Agriculture, Forest Service (Ogden, Utah).
Research Paper INT-140.

D-03-09.doc 41
EUFIRELAB
KOURTZ, P. H. and W. G. O'REGAN. 1971. A model for a
small forest fire to simulate burned and burning
areas for use in a detection model. Forest Science
17 (2): 163-169.
LEE, S. L. and H. W. EMMONS. 1961. A study of natural
convection above a line fire. Journal of Fluid
Mechanics 11 (3): 353-368.
LEE, S. L. 1972. Fire research. Applied Mechanics
Review (May): 503-509.
LOPES, A. M. G., CRUZ, M. G. and D. X. VIEGAS. 1998.
FIRESTATION – An integrated system for the
simulation of wind flow and fire spread over complex
topography. Proc. III International Conference on
Forest Fire Research / 14th Conference on Fire and
Forest Meteorology (Luso, 16/20 November 1998),
Vol. I, pp. 741-754.
LUKE, R. H. and A. G. MCARTHUR. 1978. Bushfires in
Australia. Australian Government Publishing Service
(Canberra), pp. 31-118.
MARGERIT, J. and O. SÉRO-GUILLAUME. 1998. Richards’
model, Hamilton-Jacobi equation and temperature
field equations of forest fires. Proc. III International
Conference on Forest Fire Research / 14th
Conference on Fire and Forest Meteorology (Luso,
16/20 November 1998), Vol. I, pp. 281-294.
MARTINEZ-MILLÁN, J. and S. SAURA. 1998. CARDIN 3.2:
forest fires spread and fighting simulation system.
Proc. III International Conference on Forest Fire
Research / 14th Conference on Fire and Forest
Meteorology (Luso, 16/20 November 1998), Vol. I,
pp. 907-926.
MC ALPINE, R. S. 1988. The acceleration of point
source fire to equilibrium spread. MSc. Thesis.
University of Montana (Missoula, Montana). 130 pp..
MC ALPINE, R. S., STOCKS, B. J., VAN WAGNER, C. E.,
LAWSON, B. D., ALEXANDER, M. E. and T. J. LYNHAM.
1990. Forest fire behaviour research in Canada.
Proc. International Conference on Forest Fire
Research (Coimbra, Portugal). Communication
A.02. 12 pp..
MC ARTHUR, A. G. 1966. Weather and grassland fire
behaviour. Australian Forest and Timber Bureau.
Leaflet n. 100.
MC ARTHUR, A. G. 1967. Fire behaviour in Eucalypt
forests. Australian Forest and Timber Bureau.
Leaflet n. 107.
MC ARTHUR, A. G. 1973. Forest fire danger meter, Mark
5. Forest Research Institute, Forest and Timber
Bureau (Canberra, ACT, Australia).
MC MAHON, C. K. et al. 1980. Combustion
characteristics and emissions from burning organic
soils. Proc. 73rd Annual Meeting of the Air Pollution
Control Association.
MORTON, B. R. 1965. Modelling fire plumes. Proc. 10th
Symposium on Combustion, pp. 973-982.
D-03-09.doc
NELSON, R. M. and C. W. ADKINS. 1988. A
dimensionless correlation for the spread of wind-
driven fires. Canadian Journal of Forest Research
18: 391-397.
NIELSON, H. J. and L. N. TAO. 1965. The fire plume
above a large free-burning fire. Proc. 10th
Symposium on Combustion, pp. 965-972.
PAGNI, P. J. and T. G. PETERSON. 1973. Flame spread
through porous fuels. Proc. 14th International
Symposium on Combustion, pp. 1099-1107.
PAGNI, P. J. 1975. Flame spread over thin solid fuels.
Journal of Heat Transfer 97:153-155.
PALMER, T. Y. and L. I. NORTHCUTT. 1975. Convection
columns above large experimental fires. Fire
Technology 11:111-118.
PHILPOT, C. W. 1968. Mineral content and pyrolysis of
selected plant materials. United States Department
of Agriculture, Forest Service. Research Note INT-
84. 4 pp..
PHILPOT, C. W. 1970. Influence of mineral content on
the pyrolysis of plant materials. Forest Science
16(4):461-471.
PUTNAM, A. A. 1965. A model study of wind-blown free-
burning fires. Proc. 10th Symposium on
Combustion, pp. 1039-1046.
RASBASH, D. J. and D. D. DRYSDALE. 1982.
Fundamentals of smoke combustion. Fire Safety
Journal 5:77-86.
RAY, S. R. and I. GLASSMAN. 1983. The detailed
processes involved in flame spread over solid fuels.
Combustion Science and Technology 32:33-48.
RICHARDS, G. D. 1990. An elliptical growth model of
forest fire fronts and its numerical solution.
International Journal for Numerical Methods in
Engineering 30:1133-1149.
RICHARDS, G. D. 1995. A general mathematical
framework for modelling two-dimensional wildland
fire spread. International Journal of Wildland Fire
5(2):63-72.
ROBERTS, S. 1989. A line element algorithm for curve
flow problems in the plane. Centre for Mathematical
Analysis, Australian National University.
CMAR5889. 20 pp..
ROTHERMEL, R. C. and ANDERSON, H. E. 1966. Fire
spread characteristics determined in the laboratory.
United States Department of Agriculture, Forest
Service (Ogden, Utah). Research Paper INT-30.
ROTHERMEL, R. C. 1972. A mathematical model for
predicting fire spread in wildland fuels. United
States Department of Agriculture, Forest Service
(Ogden, Utah). Research Paper INT-115. 40 pp..
42
EUFIRELAB
ROTHERMEL, R. C. 1983. How to predict the spread and
intensity of forest and range fires. United States
Department of Agriculture, Forest Service,
Intermountain Forest and Range Experiment
Station. General Technical Report INT-143. 161 pp..
ROTHERMEL, R. C. and G. C. RINEHART. 1983. Field
procedures for verification and adjustment of fire
behaviour predictions. United States Department of
Agriculture, Forest Service (Ogden, Utah). General
Technical Report INT-142.
ROTHERMEL, R. C., WILSON, R. A., MORRIS, G. A. and S.
S. SACKETT. 1986. Modelling moisture content of
fine dead wildland fuels: input to BEHAVE fire
prediction system. United States Department of
Agriculture, Forest Service. Research Paper INT-
359.
ROTHERMEL, R. C. 1990. Modelling fire behaviour. Proc.
International Conference on Forest Fire Research
(Coimbra, Portugal). Closing Conference. 19 pp..
ROTHERMEL, R. C. 1991. Predicting the behaviour and
size of crown fires in the Northern Rocky Mountains.
United States Department of Agriculture, Forest
Service, Intermountain Forest and Range
Experimental Station (Ogden, Utah). Research
Paper INT-438.
SANDHU, S. S. 1970. Mechanisms of heat transfer to
fibrous fuel ahead of an advancing flame. MSc.
Thesis. Chemical Engineering Department,
University of New Brunswick (New Brunswick,
Canada). 85 pp..
SCHROEDER, M. J. and C. C. BUCK. 1970. Fire weather -
A guide for application of meteorological information
to forest fire control operations. United States
Department of Agriculture, Forest Service.
Handbook 360.
SHAEFER, V. J. 1957. The relationship of jet streams to
forest wildfires. Journal of Forestry 55 (6): 419-425.
SIMARD, A. J. and J. E. EENIGENBURG. 1990.
METAFIRE: a system to support high level fire
management decisions. Fire Management Notes.
SMITH, R. K. 1967. Radiation effects on large fire
plumes. Proc. 11th Symposium on Combustion, pp.
507-515.
STEPHEN, C. H. and H. A. WRIGHT. 1974. Ignition
capabilities of non-flaming firebrands. Journal of
Forestry (Oct.): 646-649.
STEWARD, F. R. 1964. Linear flame heights for various
fuels. Combustion and Flame 8: 171-178.
STEWARD, F. R. 1971. A mechanistic fire spread model.
Combustion Science and Technology 4:177-….
STEWARD, F. R. and WAIBEL, R. T. 1973. Fire spread
through uniform fuel matrices. West State Section
Meeting on Combustion (Tempe, Arizona). WSC1-
73-3.
D-03-09.doc
STEWARD, F. R. 1974. Fire spread through a fuel bed.
In: Heat transfer in fires: thermophysics, social
aspects, economic impact. Blackshear, P.L. (ed.).
Ed. Scripta Book Company (Washington DC), pp.
317-378.
STOCKS, B. J., LAWSON, B. D., ALEXANDER, M. E., VAN
WAGNER, C. E., MC ALPINE, R. S., LYNHAM, T. J. and
D. E. DUBÉ. 1989. The Canadian Forest Fire Danger
Rating System: an overview. Forestry Chronicle
65:450-457.
SUSOTT, R. A. 1982a. Characterisation of the thermal
properties of forest fuels by combustible gas
analysis. Forest Science 28(2):404-420.
SUSOTT, R. A. 1982b. Differential scanning calorimetry
of forest fuels. Forest Science 28(4):839-851.
SUSOTT, R. A. 1984. Heat of preignition of three woody
fuels used in wildfire modelling research. United
States Department of Agriculture, Forest Service,
Intermountain Forest and Range Experimental
Station (Ogden, Utah). Research Note INT-342. 4
pp..
TARIFA, C. S., DEL NOTARIO, P. P. and F. G. Moreno.
1965. On the flight paths and lifetimes of burning
particles of wood. Proc. 10th Symposium on
Combustion, pp. 1021-1037.
TELISIN, H. P. 1974. Flame radiation as a mechanism of
fire spread in forests. In: Heat transfer in flames.
Afgan, N. H. and J. M. Beer (ed.). Ed. John Wiley &
Sons. pp. 441-449.
THOMAS, P. H. 1963. The size of flames from natural
fires. Proc. 9th Symposium on Combustion, pp. 844-
859.
THOMAS, P. H. 1967. Some aspects of the growth and
spread of fire in the open. Journal of Forestry
40(2):139-164.
THOMAS, P. H., SIMMS, D. L. and G. H. WRAIGHT. 1964.
Fire spread in wooden cribs. Fire Research Station
(Borehamwood, Hertsfordshire, England). Fire
Research Note No. 537. 38 pp..
THOMAS, P. H. 1971. Rates of spread of some wind-
driven fires. Journal of Forestry 44(2):155-175.
VAN WAGNER, C. E. 1964. History of a small crown fire.
Forest Chronicle 40 (2): 202-205, 209.
VAN WAGNER, C. E. 1967. Calculations on forest fire
spread by flame radiation. Canadian Forestry
Service, Department of Forestry and Rural
Development, Petawawa Forest Experiment Station
(Chalk River, Ontario). Report N.1185. 13 pp..
VAN WAGNER, C. E. 1969. A simple fire-growth model.
Forest Chronicle 45:103-104.
VAN WAGNER, C. E. 1977. Conditions for the start and
spread of crown fires. Canadian Journal of Forestry
Research 7 (1): 23-34.
43
EUFIRELAB
VAN WAGNER, C. E. 1987. Development and structure
of the Canadian Forest Fire Weather Index System.
Canadian Forest Service (Ottawa). Technical
Report 35.
VAN WAGNER, C. E. 1989. Prediction of crown fire in
conifer stands. Proc. 10th Fire and Forest
Meteorology Conference (Ottawa, 17/21 April), pp.
207-212.
VASCONCELOS, M. J. 1988. Simulation of fire behaviour
with a geographic information system. MSci. Thesis.
School of Renewable Natural Resources, University
of Arizona (Tucson, Arizona).
VASCONCELOS, M. J. P., PAÚL, J. C. U. and A.
GONÇALVES. 1998. GEOFOGO – Testing a forest
fire simulation system. Proc. III International
Conference on Forest Fire Research / 14th
Conference on Fire and Forest Meteorology (Luso,
16/20 November 1998), Vol. I, pp. 889-890.
VAZ, G. C. 1997. Prediction of the properties of a fire
front spreading through a solid porous fuelbed:
critical analysis of the state-of-the-art. MSc. Thesis
(in portuguese). Mechanical Engineering
Department, Faculty of Science and Technology,
University of Coimbra.
VAZ, G. C., ANDRÉ, J. C. S. and D. X. VIEGAS. 1998.
Proc. III International Conference on Forest Fire
Research / 14th Conference on Fire and Forest
Meteorology (Luso, 16/20 November 1998), Vol. I,
pp. 547-564.
VIEGAS, D. X., RIBEIRO, P. R. and L. MARICATO. 1998.
Proc. III International Conference on Forest Fire
Research / 14th Conference on Fire and Forest
Meteorology (Luso, 16/20 November 1998), Vol. I,
pp. 325-342.
D-03-09.doc
VOGEL, M. and F. A. WILLIAMS. 1970. Flame
propagation along matchstick arrays. Combustion
Science and Technology 1:429-436.
VON NIESSEN, W. and A. BLUMEN. 1988. Dynamic
simulation of forest fires. Canadian Journal of
Forestry Research 18: 805-812.
WADE, D. D. 1984. Fire science adaptations for the
Southern US – A research update 1980-1984. Proc.
Prescribed Fire and Smoke Management in the
South Conference (12/14 Sept., Atlanta), pp. 101-
112.
WEBER, R. O. 1989. Analytical models for fire spread
due to radiation. Combustion and Flame 78:398-
408.
WEBER, R. O. 1990. A model for fire propagation in
arrays. Mathematical and Computer Modelling
13(12):95-102.
WEBER, R. O. 1991. Modelling fire spread through fuel
beds. Progress in Energy and Combustion Science
17:67-82.
WILLIAMS, F. A. 1969. Scaling mass fires. Fire
Research Abstract and Reviews 11(1):1-23.
WILLIAMS, F. A. 1982. Urban and wildland fire
phenomenology. Progress in Energy and
Combustion Science 8:317-354.
WILSON, R. A. 1985. Observations of extinction and
marginal burning states in free burning porous fuel
beds. Combustion Science and Technology 44:179-
193.
WILSON, R. A. 1990. Re-examination of ROTHERMEL'S
fire spread equations in no wind and no slope
conditions. United States Department of Agriculture,
Forest Service (Ogden, Utah). Research Paper INT-
434. 13 pp..
44