 |
GEOGRAPHIC INFORMATIONS SYSTEMS (GIS)
Description
A geographic information system (GIS) is
a computer-based system for integrating and analyzing spatial
data. Commonly used spatial data layers in forest planning
and management include maps of vegetation type, soils, streams
and watershed boundaries, fuels, fire history, management
unit boundaries, political boundaries, land use history, and
topography. Most agencies already have data in GIS format
and progress toward the integration of databases with GIS
is ongoing.
Applicability to CRAFT
Management questions that involve large
spatial scales or spatially heterogeneous site factors lend
themselves to GIS analyses. When properly integrated, GIS
technology can form the backbone of a landscape scale risk
assessment (e.g., Burton 1999). GIS can help in the construction
of probabilistic belief nets in various ways. First, analysis
of historical patterns can be used to predict the range of
future conditions that are likely, once influential spatial
factors are identified. For example, spatial analysis of fire
severity may show that fire effects vary with topography,
and this insight can be used in spatial successional modeling.
Second, spatial uncertainties that could be important for
wildfire or wildlife habitat can be modeled using GIS. Third,
users can sample GIS layers to inform conditional probability
tables in belief networks.
Examples
- The probability that a crown fire will occur on upper slopes may differ from lower or middle slopes. These different probabilities can be calculated from fire severity assessments and a digital elevation model using GIS.
- The coincidence of fire effects and land use history can be similar explored to inform probability tables
- The detailed attributes obtained from a sample of vegetation plots or FIA points can be probabilistically assigned to vegetation strata layers
- The density or absence/presence of a wildlife species may vary with vegetation type or topography. With enough point data, GIS can be used to assign probabilities to unsampled areas with similar environmental attributes.
- The range of attributes associated with forest
patch sizes could be analyzed in GIS.
Back to Top
|
FARSITE
Description
 FARSITE
is a landscape fire model that simulates fire growth in the
landscape using GIS layers. It integrates place-specific vegetation,
fuel and topographic attributes with streaming weather data
to predict attributes of fire behavior including flame length,
spotting distance and rate of spread (Finney and Andrews 1999).
FARSITE requires five GIS layers: elevation, slope, aspect,
fuel model and canopy cover. Additional layers that may be
important depending on the question asked include tree height,
crown base height, crown bulk density, duff loading and woody
surface fuel.
Applicability to CRAFT
FARSITE is typically used to address how
fire spreads through the landscape and with what behavior.
Fire does not spread through mountainous landscapes uniformly,
but can be highly sensitive to local winds and slope. Given
a FARSITE landscape, it is easy to calculate likely weather
parameters and possible ignition locations using probability,
but risk assessments should be informed by more than a few
weather or ignition scenarios. A manager could run fire spread
with multiple weather scenarios and ignition locations, but
this effort may prove to be very time consuming for large
areas. It is the range of possible outcomes and their likelihoods
that are needed to inform conditional probability tables in
CRAFT. Such modeling is more feasible with small areas with
concerns about fire spread from a single source area, such
as areas with a high incidence of arson. Recent work to identify
the most likely routes of fire spread in the landscape using
FARSITE could be used to address the suppressibility of fire
in landscapes managed according to different treatment options.
For most landscape scale characterizations of risk, users
may find FLAMMAP to be more applicable (see section below).
The key data uncertainties that FARSITE modelers
face vary place to place, but some uncertainty is present
in every FARSITE modeling effort given the heterogeneity of
fuel and vegetation. Fuel conditions are especially incomplete
in GIS. Classification of pixels according to their fuel model,
and attribution of canopy cover, canopy height, canopy bulk
density and canopy base height introduces plenty of spatial
uncertainty because some of these variables are derived from
limited field sampling, some are based on remote sensing with
limited ground confirmation, and others are derived or inferred
using expert opinion. In contrast, elevation, slope and aspect
derived from Digital Elevation Models (DEMs) are accurate
layers.
Examples
- The likelihood that fire that starts from a specific ignition point will reach a specific destination, such as the wildland urban interface, can be modeled using FARSITE. Conditional probability tables might include a range of arrival times or a range of possible weather scenarios.
- The effectiveness of shaded fuel breaks
on landscape fire severity patterns can be probabilistically
addressed using a range of fuel break widths, locations,
or treatments. To be useful, a wide range of fire weather
scenarios would be required and multiple start locations
would further increase the reliability of results. Due to
the large number of combinations, this would involve substantial
model runs.
- Managers interested in modeling fire extent or severities could do so using a range of fire weather scenarios. To be useful, most risk assessments would also require multiple ignition sites and this would result in considerable map output.
Website: FARSITE
Back to Top
|
FLAMMAP
Description
FlamMap generates maps of fire behavior (such as rate of spread, crowning, flame length, etc.) by integrating GIS layers of fuel and vegetation with topography and fixed weather conditions. Unlike FARSITE, FlamMap does not calculate fire spread or differential behavior that results from header or flanking fires. Neither does it use streaming weather data. Instead, it models fire behavior that is likely to occur at different locations, given topography, fuels and specified weather station parameters.
Applicability to CRAFT
The advantage that FlamMap has over FARSITE
is that it is much easier to control parameters while changing
others. This characteristic greatly facilitates the acquisition
of data for filling out conditional probability tables in
CRAFT. This allows more user control over integrating FlamMap
into a belief network.
For most fire or fuels questions, the specific
outcome of an individual fire under one fire spread scenario
is less important than the most likely effects. Although the
most likely effects could be calculated using multiple FARSITE
runs, it would be time consuming to do so. As an alternative
approach, FlamMap allows users to ignore the nuances of fire
spread and hold fire weather constant. FlamMap is useful for
landscape questions because it performs similar calculations
to what can be obtained from BehavePlus for a single stand,
but for the entire landscape at once. Such landscape scale
questions comprise an important part of cumulative effects
analyses.
Examples
- The range of flame lengths expected across
a heterogeneous burn unit, watershed, wildfire perimeter
or along a fuel break can be calculated based on fire weather
conditions (e.g., percentile fire weather)..
- Fuel conditions can be varied based on
the range of conditions that may be possible. While the
actual fuel conditions are known for only a fraction of
the landscape, assigning the range of possible conditions
will allow for sensitivity assessments once the belief network
is constructed. This range of fuel conditions could be assigned
probabilistically using FIA sample data from similar vegetation
types.
- In combination with GIS analysis, the likelihood of a given flame length, spread rate or spotting distance can be calculated according to topographic variables, vegetation type, etc.
Website: FlamMap
Back to Top
|
INFORMS-MSN
(Most Similar Neighbor)
Description
Most Similar Analysis (MSN) is an INFORMS
application that decreases the uncertainty of spatial vegetation
characteristics within a landscape. Uncertainty may exist
due to low sampling density and it increases with time since
the last stand exam. Landscape scale analyses performed using
GIS, FARSITE or FlamMap often require information about portions
of the landscape that have no inventoried data. When variables
of interest involve fuels or vegetation, uncertainty can have
a substantial effect on the reliability of fire behavioral
output.
MSN uses the information that is available
for a portion of the landscape in conjunction with topographic
information (i.e., a DEM) and remote sensing to assign attributes
for missing portions of the landscape based on their similarities.
This process assigns values based on the attributes that unknown
areas are most likely to have. As long as spatial correlations
hold true, MSN can improve the quality of GIS layers.
Applicability to CRAFT
MSN can be a valuable tool for deriving
a more accurate spatial data layer for use in FARSITE or FlamMap.
Given its current capabilities, it is most likely to be of
value for CRAFT as a supportive tool for FARSITE or FlamMap
analysis rather than as a stand alone resource. To be most
applicable for probability analyses, future versions of MSN
would create a distribution of states for an unknown polygon based
on probabilistic analysis. This could make many uncertainties
of the MSN process more transparent.
Examples
- Use MSN to predict fire behavior in FARSITE
or FlamMap over less accurate maps if a landscape characteristic
is highly sensitive to different fuel or vegetation parameters
in your belief network sensitivity analysis.
Website: INFORMS
Back to Top
|
BEHAVE PLUS
Description
BehavePlus is software that models fire
behavior (i.e., flame length, rate of spread and spotting
distance, etc.) based on data for a single place and point
in time. BehavePlus is the standard software used to model
surface fire behavior. Although BehavePlus and FARSITE/FlamMap
use many of the same equations, BehavePlus was not designed
to address landscape scale fire behavior. Newer versions of
BehavePlus include a tree mortality component that matches
those equations used in FOFEM. With continued development,
BehavePlus will integrate a range of other programs (Andrews
and Bevins 1999). Among the weakest components of this or
any existing fire behavior software is how crown fires are
modeled. Crowning is sensitive to live ladder fuels which
is often lacking and highly variable in many forests.
Applicability to CRAFT
As a stand-alone model, BehavePlus is useful
for management questions that address stand-level rather than
landscape level questions of fire risk. Long-term successional
changes in fuel could be modeled separately and used as inputs
in BehavePlus and site parameters could be varied to reflect
different classed portions of the landscape. For broader questions,
local controls on wildfire behavior could be modeled for classed
portions of a watershed and then assigned probabilistically
in GIS. By varying one or more parameters in BehavePlus, CRAFT
users can efficiently define a range of expected values to
use in belief net conditional probability tables.
Examples
- To determine possible variation in flame lengths within an area to be prescribed burned, calculate how flame lengths respond to variation in fuel moisture and fire weather
- Calculate change in the rate of spread to variation in wind direction based on historical data
- To determine fuel break effectiveness, calculate a range of spotting distances in a stand that result from different flame lengths and spotting tree species.
- To address the effects of uncertain fuel conditions, calculate flame lengths with different custom fuel models
Website: BehavePlus
Back to Top
|
NEXUS
Description
NEXUS is software designed to model crown fire potential
(Scott 1999). NEXUS links surface and crown fire models to
calculate crown fire hazard within a single stand using fuel
models, fuel moisture, canopy fuels, wind and topography as
inputs.
Applicability to CRAFT
The initiation and spread of crown fires are poorly understood compared to surface fire behavior. This uncertainty is partially due to the complex variability of wind and fuel, limited direct observational data due to risks to researchers, and the complexities involved with modeling thresholds of crown fire initiation. Given these inherent uncertainties that surround crowning, management questions that involve crown fire risk lend themselves to probabilistic analyses in CRAFT. As with other fire behavior models, specifying different values for fuel and fire weather provides a range of data that can be used in belief networks. The likelihood that these varied factors actually occur within the plot or for a specified burn period can be obtained through frequency analysis of sampled plot data or historical weather data.
Examples
- Calculate crown fire hazards under a range of fuel moisture and wind conditions
- Determine fireline intensities using a range of dead fuel moisture conditions or canopy base height values
Website: NEXUS
Back to Top
|
CONSUME
Description
Consume predicts the consumption and emissions from dead fuels based on weather and fuel moisture. Consume was designed to support fuel treatments after logging, pile burning and natural fuels. It lends itself, therefore, to analysis of fuel consumption and emissions that can be expected using different management alternatives.
Applicability to CRAFT
As with other fire-related models, the usefulness of the output provided by Consume reflects the assumptions of the model, the quality of the data entered, and how users apply the model to obtain data for use in CRAFT. CRAFT users can use Consume in much the same way as that suggested for BehavePlus and FOFEM. The equations in all of these programs result in a single set of output in response to specified input. The different variables to alter in CONSUME depend on the data needs of the belief network in CRAFT, and the variables of interest and those that have the most uncertainty. More than other parameters, fuels may be extremely variable and of unknown conditions through a burn unit and fuels can also vary markedly over time.
Examples
- Document how fuel consumption differs in response to variable fuel moisture or humidity
- Determine variation in the smoke emissions that could result from different proportions of sound versus rotten fuel.
- Consider the flammability of fuel in
response to a range of times between harvesting and fuel
treatment.
Website: CONSUME
Back to Top
|
FIRE FAMILY PLUS
Description
FireFamily Plus is software for analysis
of a wide range of daily RAWS weather data and fire indices.
Users can calculate fire weather conditions from all or a
specified portion of the historical record from one or more
stations. FireFamily Plus also allows users to analyze historical
fire data in conjunction with fire climate data.
Applicability to CRAFT
FireFamily Plus provides an efficient means
to collate, sort and analyze historical weather data for use
in CRAFT. Weather data can be exported from FireFamily Plus
for use in FARSITE and the values calculated in this software
can be used as input into other software, such as BehavePlus,
Consume or FlamMap. For example, FireFamily Plus can be used
to determine what specific fuel moisture or fire weather values
constitute extreme, high, moderate or low conditions, and
how likely they are to occur, given the period of record.
Examples
- Determine the specific fire weather values
that constitute the 97th , 90th, and 75th percentile conditions
(e.g., by using Burning Index) to be used in multiple BehavePlus
runs
- Specify streaming weather conditions
to be used as input into FARSITE
- Calculate mean, maximum or minimum monthly
weather values for a spring versus a fall prescribed burn
- Calculate the range and distribution
of weather or fuel moisture conditions that are likely to
occur during a specific week or month of the year
- Determine the
likely date of a season-ending precipitation event
Website: FireFamily Plus
Website: FireFamily Plus User's Guide - RMRS-GTR67www
Back to Top
|
RERAP
Description
The Rare Event Risk Assessment Process (RERAP)
calculates the probability that fire will spread to undesired
locations or that a fire-ending precipitation event will occur.
RERAP is capable of addressing landscape scale questions,
but it does so differently; that is, not with the three-dimentional
spatial knowledge that is used in FARSITE or FlamMap. As a
result, the limiting assumptions of this software are that
fire will actually arrive at a destination along the given
pathway according to fire weather. Fire spread often follows
preferred topographic routes that RERAP can not address (this
is better modeled with FARSITE). For example, fire spread
would occur much faster than modeled if complex topogrpahic
patterns favor fire spread. Fire spreads uphill rapidly and
crowning behavior on upper slopes may be important factor
governing the rate of spread.
Rates of spread observed on the surface
are often less than that from crowning and spotting. RERAP
has no internal mechanism for modeling crown fire initiation
or spread, instead, this work is performed in Nexus, and to
some degree, BehavePlus (used in the USA) or the Fire Behavior
Prediction Model (used in Canada). As a result, the values
obtained for rates of spread are dependent on the limited
ability of these programs to model crown fires. RERAP uses
climatological data exported from FireFamily Plus.
Applicability to CRAFT
RERAP provides a venue for calculating specific
probabilistic parameters that can be input directly into conditional
probability tables in CRAFT. In other words, RERAP can be
used as a vehicle for defining probabilistic relationships.
Users should be aware of its assumptions, when addressing
questions that involve complex spatial patterns.
Examples
- Use RERAP to calculate the probability that a fire or smoke from a single location will reach a second location
- Describe the likelihood of a fire-ending storm event for a given date
- Determine the length of the fire season or the burning time probable, given different ignition dates
Website: RERAP
Key references:
Wiitala MR, Carlton DW (1994) Assessing long-term fire movement risk in wilderness fire management. In ‘Proceedings of the 12th conference on fire and forest meteorology, 26–28 October 1993, Jekyll Island, GA’. pp. 187–194. (Society of American Foresters: Bethesda, MD)
RERAP User's Guide: Rare Event Risk Assessment Process, Version 5.03, (2000) USDA Forest Service
Back to Top
|
SUPPORTIVE LINKS
FRAMES Links to fire models
Fire.org Links to fire models
|
|
![[Text graphic] Fire Models and Tools](../assets/Titles_HumanstFont/Title_FireModelsTools.gif)
The
following fire-related models are among the most frequently
used to predict fire behavior, hazards, fire weather and fire
effects.