Fire Behavior

1
Fire Behavior
Fire spread and intensity Key interrelated
aspects of fire behavior are Rate of spread
(ft/min, ch/h, m/h) Fireline intensity
(heat/area/time ? BTU/ft2/sec) Heat per unit
area (BTU/ft2).
2
Rate of spread (ROS) is movement of the fire
perpendicular to the fireline.
3
Rate of spread can vary, usually consider an
average. ROS is fastest at the head of the
fire (driven by wind or upslope), least at back,
intermediate at flanks. Note shift in winds or
slope can cause heading ROS to shift within fire
(flanks can become head). The fuels with the
greatest effect on ROS are fine fuels needles,
grass, leaves, twigs. ROS can be measured in
the field best to do so with watch/timer and
known reference points, such as measured pinflags
in research burns. ROS of high-intensity, fast
wildfires can be measured from the air, mapped.
4
Intensity is heat (energy) release per unit area,
proportional to flame length. FL 0.45
IB0.46 (English units, ft BTU/ft/sec) Note
distinguish flame length from flame height. Why
is flame length useful? Easily observed in the
fieldbut note that it is can be variable
subjective. Try to use objects of known height
as reference points, try to estimate averages and
maximum FL. IB Byrams fireline intensity
(heat released per second from a foot-wide
section of fuel extending from the front to the
back of the flaming zone.
5
(No Transcript)
6
Photo by Brian Conway, Bconway1_at_home.com
7
How much heat does a given area receive?
Reaction intensity times residence time.
Residence time is inversely proportional to
surface-area-to-volume. Recall fine fuels have
high surface-area-to-volume ratio residence time
is low in fine fuels ? less heat per unit area
even at the same intensity.
8
Fire Characteristics Chart
9
Fires with identical intensity (flame length) can
have very different heat output, thus strikingly
different effects. The example fires in
Andrews figure are a fast-spreading fire (A) vs.
a slow fire (B).
10
Limitations of Fire Behavior Modeling
BEHAVEand the thrust of most of the entire fire
behavior modeling effortare oriented toward
prediction of the characteristics of the flaming
front. This information is most relevant to fire
assessment for suppression, prescribed burning,
etc. Useful for some fire effects such as crown
scorch flame length and fireline intensity are
best related to the effect of fire on items in
the flame and in the hot convective gases above
the flame Andrews 198657). Key point
description of the flaming front characteristics
is incomplete for evaluating fire effects,
especially with respect to smoldering fires and
to combustion of large fuels behind the flaming
front. Note Andrews point (p. 63) that fire
behavior must be quantified, cannot just compare
fire/no-fire or hot and cool for
understanding effects or ecology.
11
Extreme fire behavior occurs on a small
percentage of fires but has major effects (in
western US 1 of the largest fires accounted for
80-96 of area burned Cited in Rothermel 1991).
Crown fire behavior is a major contributor to
tree mortality in forest fires. Crown fire can
be a natural ecological disturbance pattern in
many forest ecosystems (lodgepole pine, boreal
forests) can represent unnatural disruption
(ponderosa pine). Whenever the goal is
suppression of a fire, crown fire makes
suppression efforts much more difficult.
12
Rothermel 1991
13
Crown fire burns through tree canopies, can be
very intense and fast. Crown fires develop due
to intense surface fires (hot, dry, windy
conditions) and canopy characteristics conducive
to overstory burning. C.E. Van Wagner (leading
crownfire researcher in Canada) identified
critical characteristics of height of the base of
the crown (note this will be an important
variable in Farsite) and foliar moisture content.
Crown fires are virtually always dependent on
surface fire, but active crown fires (spreading
with the surface fire) and independent crown
fires have been identified (Van Wagners
terminology). Torching or passive crown fire
is a pre-crownfire phenomenon, occurring when
canopies of one or a group of trees burn
14
(No Transcript)
15
Spotting by firebrands can occur as burning
embers are lifted in the convective column of any
fire the stronger and higher the column, as in a
torching or crowning fire, the bigger embers can
be lifted and the further they can be carried.
Embers can be made of a variety of fuels
cones, grass, moss, wood. Bigger embers
usually travel shorter distances, but are more
likely to stay ignited than smaller embers.
The probability of ignition of spot fires
depends on the fuel type and fuel state where the
ember lands (e.g., contrast cured vs. green
grass, rotten vs. sound wood, dry vs. moist
weather). Spots can start new fires miles
ahead of the flaming front in severe situations.
Spot fires are often a threat to fire control in
both wildfire and prescribed fire situations.
16
Fire whirl is a vortex, rotating gas mass. Can
become strong enough to break 3 dbh trees, can
lift firebrands, uproot shrubs. Most are small.
Usually develop near ground, like a dust devil,
but sometimes develop above ground and extend
down, like a tornado. Fire whirls arise from
convection columns when rotational motion occurs,
possibly due to friction from an obstruction
(e.g., tree canopy). Often appear where eddies
occur in the airflow, like ridgetops, canyon
bends. Unstable atmosphere favors firewhirl
development.
17
Effects of firewhirls (1) increase combustion
rate as surface air is drawn in ? increased fire
intensity, more complete combustion (2) pick up
bigger firebrands and scatter them further and
faster--fire whirl can move at the speed of the
prevailing wind (while fire fronts usually move
much slower) Horizontal roll vortex is a very
extreme event where rotating air moves
horizontally, driven by strong winds. Andrews
describes unburned crown strips or streets,
believed to be caused by cooler air drawn into
the vortex, in the Mack Lake, MI, fire in 1980.
18
Predicting fire behavior Surface fires
(Rothermel/BEHAVE/Canadian FBPS) Experience
and knowledge of fire managers has always been at
the heart of predicting fire behavior and
interpreting fuel and weather conditions.
Predicting fire behavior in order to support
fire management decisions was the major
motivation for fire behavior research. Models
do not replace experience, but they can offer
decision support, they are quantitaive and
repeatable, give insight into natural mechanisms,
and a basis for training and further
experimentation.
19
The fire behavior prediction model used in the US
is Rothermels model (based on work by many
people). How is it used? Not by direct
mathematical solutioninstead, simplifications
fuel models are used (1) model relationships
are expressed as graphic aids, especially
nomograms. Can be summarized very succinctly
(example fireline handbook). (2) computerized
versions of the fire behavior model, first for a
handheld calculator with a special chip (TI-59),
then BEHAVE program, then FARSITE NEXUS.
20
Who makes fire behavior predictions? Everyone
associated with fire decisions, especially
initial attack firefighters. Predictions may
take a very simple, qualitative form, but are
inherently part of the firefighting (or
prescribed burning) process. Firefighters must
be aware of effects of slope, weather changes,
winds, moisture, fuels, etc. On large
wildfires, specialized Fire Behavior Analysts are
responsible for predictions. They draw on
special fire weather forecasts, reconnaissance of
the fire, aerial surveys, and fuel data. Their
responsibility is not only to develop numerical
predictions, but to help interpret these results
in terms of fireline perimeters at different
points in time, predicted fire runs, etc. In
planning for prescribed fires, managers develop
predictions to forecast the suitability of fire
effects.
21
Limitations on prediction Variability in fuels,
topography, and weather is the major limitation
to fire behavior predictions. Wildland fuels are
exceptionally complex (variety of chemical
constituents, complex patterns of vegetation
growth and death). Usually fuel information is
limited to general fuel model description. There
are only 13 standard fire behavior fuel models
covering the entire US, so fire behavior
predictions are necessarily simplified.
Example fuel model 9 (hardwood litter)
covers forest types from oak-hickory to ponderosa
pine. The fire behavior model is
deterministic, so a given set of inputs always
results in the same output. So identical model
predictions may need to be applied to fairly
different situations. Wherever fuel inventory
information exists, it is possible to develop
custom fuel models. Even when the data is good,
however, there are still usually differences from
place to place, different moisture (e.g., due to
shading).
22
Limitations on prediction (cont.) Topographic
differences strongly affect small- and
large-scale fire behavior, but often modelers
have to deal with average slopes
aspects. Limitations on weather forecasting are
probably the greatest obstacle to successful
long-term fire behavior predictions (e.g.,
several hours to days). Whenever the weather
pattern is consistent, modeling works well and
modelers can plan for expected diurnal (daily)
changes, such as the shift to downslope winds in
the evening. But weather patterns can change
abruptly with significant effects on fire
behavior, or weather patterns may exceed the
presently-expected limits of probability.
Example severity of 1988 Yellowstone drought,
1996 2002 SW droughts ? events that exceed our
historical record confound our planning.
23
Crown fire prediction Crown fires present the
greatest threat to resources and the greatest
hazard to suppression forces, but crown fire
behavior is the most difficult to predict. Fire
behavior prediction models such as Rothermel's
apply to surface fires. Why are crown fires
harder to model? Effects of variability in
fuels, topography, weather are even more
important, crown fuels can be discontinuous,
harder to build up the heat required to get fire
into crowns.
24

• (Following discussion from Rothermel 1991)
• Transition from surface to crown fire is
  facilitated by several conditions
• dry fuels
• low humidity and high temperatures
• heavy accumulations of dead and downed litter
• conifer reproduction and other ladder fuels
• steep slope
• strong wind
• unstable atmosphere
• continuous forest of conifer trees

25
Forest structure influences fire behavior
Crown bulk density
Fuel model 2 or 9
Canopy base height
Fuel model 9 or 10
26
Wind-driven crown fire is supported by strong
winds (ROS 1-7 miles/hour). Running crownfire
has showers of sparks and embers, spotting ahead
of the flaming front, smoke, strong convection
column. Plume-dominated crownfire develops a
very strong convection column that accentuates
atmospheric instability, creates turbulent
surface winds. Plume-dominated crownfires can be
enhanced by reverse wind profile (winds are
faster at surface than higher up, so convection
column is not sheared away by winds
aloft). Rothermel (1991) developed preliminary
fire behavior prediction techniques for
wind-driven crownfires, empirically relating
predicted surface ROS to observed ROS on several
large fires. The average rate spread for
crownfires was 3.3 times faster than predicted
surface ROS. Developed fire characteristics
charts nomograms for crownfires. Finney
modeled crownfire behavior in an interactive fire
behavior model, FARSITE.
27
(No Transcript)
28
Raster Landscape Themes for FARSITE Simulations
220
210
200
200
205
205
Elevation
205
220
210
205
200
200
200
200
200
205
250
205
230
205
210
205
200
200
Slope
200
200
205
200
205
220
240
210
205
205
200
200
200
200
Aspect
205
220
210
190
205
240
210
200
205
200
200
205
200
20
210
Fuel Model
190
30
210
220
5
205
3
200
10
200
7
190
Canopy Cover
110
3
110
5
180
2
270
2
270
270
5
170
4
5
Canopy Height
14
170
130
10
190
10
10
200
10
120
0
10
Crown Base Height
0
140
2
80
6
80
60
6
2
60
0
60
2
0
Crown Bulk Density
20
20
0
10
0
25
25
20
0
25
0
20
5
2.0
0
0.4
0.4
0
1.2
15
0
1.2
0
15
0.2
,23
0
.25
.25
0
.28
0.2
.28
0.2
.05
0
0
.05
.05