Landscape Hazard Assessment Past Approaches and Current Modeling Tools.

Landscape Hazard Assessment

Past Approaches andCurrent Modeling Tools

• Vegetation/Fuels Information was ground truthed.• Rating Included Ignition

Likelihood, Values at Risk, and Suppression Difficulty• Fuels Hazard based on static

evaluation of each location. Not able to evaluate change based on Fuel Treatments

• Fire behavior calculated independently for each pixel

• Used fireline intensity (analog for flame length)

• No assessment of ignition or spread

Adding Fire Growth PotentialFLAMMAP Burn Probability/IFT-DSS RanDig

FLAMMAP 5 is stand alone software, with latest version of the Minimum Travel Time (MTT) fire growth model for Burn Probability and Treatment Optimization calculations.• It includes spotting as spread vector• Calculation intensive analyses like

burn probability can take hours to complete

IFT-DSS is online tool, in development. • It includes FLAMMAP

3, which does not incorporate spotting spread

• Calculations made on server, saving local computer resources

Factor Issues forBurn Probability Analysis

• What are vectors of spread• What environmental scenarios are of interest• Ignition frequency and distribution• What are barriers• How are fuel treatments and fire scars

represented• Are you evaluating individual landscape

changes or evaluating hazard in a general sense• Size of Analysis Area

Effect of Landscape Classification

LANDFIRE 2008

CWPP 2006

LANDFIRE Landscapes are consistently developed across the entire US. Limited use of local mapping of vegetation. Significant edits are normally required for analysis.

Locally produced versions, like this CWPP project map may be more accurate. They are often limited by ownership boundary, image availability, and the time for production.

Is this Masticated Fuel Break a Barrier?

What do you think now?

Fire Behavior in Treatments & Burn Scars

Effect of Ignition Source

Random IgnitionsAnd

Spatially Explicit Ignitions

Lightning and Human Ignitions• 2013 Ignition Cause

402 -- Human211 -- Lightning

• 2014 Ignition Cause339 -- Human54 -- Lightning

Lowery, James, "The effect of random and spatially explicit lightning and human-caused ignitions on simulated burn probabilities at small scales" (2012). Theses, Dissertations, Professional Papers. Paper 992

10,000 Random Ignitions 3,500 Human Ignitions

Effect of Environment Factors

Fuel MoistureWindspeed

Wind Direction

WindspeedWind Direction

Wind-Driven Surface Fires

DMC Driven

Cumulative Drought

StageDiurnal Effect

Stage

How the FLAMMAP Burn Probability Model Works

10,000 overlappingFire Perimeters

Before Recent Fires

10,000 overlappingFire Perimeters

After Recent Fires

The Inputs Matter!

The Results

Burn ProbabilityConditional Flame Length

Fire Size Distribution

Burn Probability Output Use enough ignitions to burn entire landscape

Conditional Flame Length Averages all flame length

from fires that burned each pixel

Evaluating Treatments• Landscape accuracy critical• Worst case environment?• Multiple wind directions?

Reprocessing Model Outputs• Combining burn

probabilities from multiple analyses, accounting for– wind direction– Landscape

changes• Other Examples

Ager, Alan A.; Vaillant, Nicole M.; Finney, Mark A.; Preisler, Haiganoush K. 2012. Analyzing wildfire exposure and source–sink relationships on a fire prone forest landscape. Forest Ecology and Management. 267: 271–283.

Concluding thoughts and questions