Comparison of Fire Behavior Modeling Methods in IFTDSS

In IFTDSS, there are currently three models used to approximate fire behavior. Each of these models has unique strengths that help answer specific questions and meet specific objectives.

Landscape Fire Behavior
MTT Fire Spread
Landscape Burn Probability

Modeling in IFTDSS compared to other applications

Each of the above models comes from FlamMap , though you may also be familiar with them through using Wildland Fire Decision Support System (WFDSS). While the fire behavior calculations from these models are nearly identical in WFDSS and IFTDSS, they are named differently in each application and may have different fixed inputs or have slightly different assumptions based on how the developers built them into the application. If you are unsure, we recommend consulting the documentation in the application you are using for clarification. Each modeling section in the IFTDSS Help Center includes a Technical Documentation topic, such as this one for Landscape Fire Behavior.

Below is a table outlining the different naming conventions in IFTDSS, WFDSS, and FlamMap 6.1. For context, we’ve also included some other popular fire behavior modeling tools that may be included in one application but not the other.

Model names and types.
Model Name	Type	Name in FlamMap	Name in IFTDSS	Name in WFDSS
FlamMap	Static fire behavior	FlamMap	Landscape Fire Behavior	Basic
Minimum Travel Time (MTT)	Fire spread	MTT	MTT Fire Spread	Short-Term Fire Behavior
Random Ignitions (Randig)	Burn probability	MTT	Landscape Burn Probability	N/A
Fire Area Simulator (FARSITE)	Fire spread	FARSITE	N/A	Near-Term Fire Behavior
Fire Spread Probability (FSPro)	Burn probability	N/A	N/A	FSPro
Large Fire Simulator (FSim)	Burn probability	N/A	N/A	N/A

Not all applications have the same models. For example, FlamMap 6.1 and IFTDSS do not include FSPro while WFDSS does. Additionally, some applications may have more than one choice for a given model type. In FlamMap for example, fire spread can be calculated using either FARSITE or Minimum Travel Time.

Choosing a model

Choosing which model to use depends upon the question(s) you are trying to answer. Some questions may best be addressed by a single model while others could be answered using one of several models.

Three fire behavior models available in IFTDSS. Under each model are example questions that the model is well suited to address.
Landscape Fire Behavior	MTT Fire Spread	Landscape Burn Probability
How effective are my proposed fuel treatments at altering fire behavior?
Which areas on my landscape are more likely to experience crown fire versus surface fire?	When will fire likely impact an area?	Which areas on the landscape are more likely to burn than others?
Which areas have the fastest spreading fuels?	How quickly can I expect fire to move across my landscape?	How likely are fires on my ownership immediately adjacent ownerships and inholdings?
Where in my burn unit might flame lengths and rate of spread exceed capability of proposed holding resources?	What are the expected rates of spread or fireline intensity due to fire from a given ignition over the burned area?	Which areas will burn with the highest intensity and severity?
	Did my treatments change the spread of fire? Slow it down or keep it out to allow for better ingress/egress?	Which areas on this landscape would benefit from treatments with objectives of reducing fire intensity in areas that are most likely to burn?

Different Modeling Approaches

The three models approach fire behavior modeling differently. For example,Landscape Fire Behavior maps fire behavior for a “snapshot" in time. It models head fire only and fire does not spread across the landscape – every pixel burns and every pixel is modeled independently, ignoring neighboring pixels.

In contrast, MTT Fire Spread and Landscape Burn Probability spread fire across the landscape from ignition(s). These models include heading, backing and flanking fire, which can result in lower intensity than Landscape Fire Behavior. Where these models differ is that MTT Fire Spread requires you to specify an ignition location and the model simulates a single fire from that location. The Landscape Burn Probability simulates thousands of fires from random point ignitions assigned by the model. In addition, MTT Fire Spread can be simulated for up to 5 burn periods; Landscape Burn Probability uses a single burn period.

Differences in model approaches to fire spread, spotting, behavior, and outputs are outlined in the table below.
	Landscape Fire Behavior	MTT Fire Spread	Landscape Burn Probability
Weather conditions	Weather conditions remain constant throughout the entire simulation time.
Types of fire spread modeled	Head fire	Head, flanking and backing fire.
Predicted Fire Behavior	Maximum	Based on type of fire spread (head, flank or backing fire).	Probabilistic flame length based on the mean intensity of all the times it burned.
Fire Spread	Fire behavior calculated independently for each pixel.	Fire spread is modeled across pixels based on Minimum Travel Time algorithm.
Spotting	Not modeled.	Modeled and influences fire spread.
Output Type	Deterministic – based on set inputs that you specify.	Deterministic – based on set inputs that you specify, spotting is stochastic, ignition is fixed.	Probabilistic – based on set inputs that you specify, spotting is stochastic, ignitions are random. Outputs are conditional probabilities based on thousands of simulated fires.

Different Model Outputs

Each model offers unique outputs to support key analysis questions or decision making. The table below shows a list of outputs that can be viewed in the Map Studio (outputs are available for download as well).

A comparison of model outputs.
Landscape Fire Behavior	MTT Fire Spread	Landscape Burn Probability
Flame Length (ft)	N/A	Conditional Flame Length (ft)
Rate of Spread (chains/hr)	MTT Rate of Spread (chains/hr)	N/A
Fireline Intensity (BTU/ft-sec)	MTT Fireline Intensity (BTU/ft-sec)	N/A
Heat per Unit Area (BTU/ft²)	N/A	N/A
Crown Fire Activity	N/A	N/A
N/A	MTT Arrival Time	N/A
N/A	MTT Major Paths	N/A
N/A	N/A	Burn Probability
N/A	N/A	Integrated Hazard

It is important to note the distinction between fire behavior outputs that, at first glance, may seem similar. For example, both the Landscape Fire Behavior model and the MTT Fire Spread model have a rate of spread output, yet they are not the same. Similarly, Flame Length and Conditional Flame Length from Landscape Fire Behavior and Landscape Burn Probability differ. These differences are due to differing model approaches discussed above. The subsections below explain these different examples in greater detail.

Rate of Spread versus MTT Rate of Spread

The example below maps rate of spread from Landscape Fire Behavior. Both simulations were completed on the same landscape using the same inputs for winds, crown fire inputs, initial fuel moistures, and fuel moisture conditioning. For MTT Fire Spread an ignition was also needed along with Simulation Time and Spotting.

Three things to notice when looking at these outputs:

The Landscape Fire Behavior output has values mapped for the entire extent. Whereas, MTT Fire Spread only has values mapped within the burned area.
Although the values are binned the same way and use the same colors, the names of the layers are different. This different naming convention was chosen to help you differentiate between similar ouputs from different models.
Pixel-per-pixel, rate of spread from Landscape Fire Behavior is usually higher than that from MTT Fire Spread. This is because Landscape Fire Behavior Rate of Spread is the maximum rate of spread (head fire), and MTT Fire Spread Rate of Spread is the based on the direction of the spread, which includes head, back and flanking fire.

Landscape Fire Behavior outputs covering the entire landscape on the left and MTT Fire Spread outputs covering only the area of fire spread on the right. Rate of spread tends to be higher in the mapped outputs on the left because Landscape Fire Behavior models only head fire.

Fireline Intensity versus MTT Fireline Intensity

This example for fireline intensity uses the same model inputs as the Rate of Spread example above. You’ll notice the same differences as above, which exist for the same reasons:

Full coverage extent for Landscape Fire Behavior versus a specific burn area for MTT Fire Spread.
Similar binning and color scheme but with a different naming convention.
Pixel-per-pixel, less fire line intensity in the MTT Fire Spread output due to its inclusion of head, back and flanking fire.

Landscape Fire Behavior outputs covering the entire landscape on the left and MTT Fire Spread outputs covering only the area of fire spread on the right. Fireline Intensity tends to be higher in the mapped outputs on the left because Landscape Fire Behavior models only head fire.

Flame Length versus Conditional Flame Length

The example below shows maps of flame length from Landscape Fire Behavior and Landscape Burn Probability. Both simulations were competed on the same landscape using the same inputs for winds, crown fire inputs, initial fuel moistures, and fuel moisture conditioning. For Landscape Burn Probability, Simulation Time and Spotting were also needed and ignitions are randomly applied by the model.

Two things to notice when looking at these outputs.

The bin classes, in addition to the colors, are different between the two models to help you distinguish that they represent two different types of flame length outputs.
Pixel-per-pixel, flame length from Landscape Fire Behavior will be higher than that of Landscape Burn Probability. This is because Flame Length is from head fire, and Conditional Flame Length is a mean value the based on the probability of burning at a given intensity which includes the direction of the spread (head, back and flanking fire).

Flame length from Landscape Fire Behavior on the left and conditional flame length from Landscape Burn Probability on the right. Flame lengths tend to be lower on the right because Landscape Burn Probability uses a mean value the based on the probability of burning at a given intensity which includes the direction of the spread (head, back and flanking fire).