Time Until Fire Arrival (TUFA) modeling Mark McLean, Ph.D. October 9, 2013
Nov 28, 2014
Time Until Fire Arrival (TUFA) modeling
Mark McLean, Ph.D. October 9, 2013
Anchor Point Company Profile
• Wildfire hazard and risk mitigation – solutions from pre-planning through mitigation management
• Company formed by firefighters • Clients both private and government
– Other consulting firms – Home owner associations – Fire departments – Local, State and Federal Government
• Hazard and risk assessments • Community Wildfire Protection Plans (CWPPs) • Fuel treatment project management • National Hazard and Risk Model (No-HARM)
Background
• Important for many reasons to know when a potential fire will arrive – Suppression trigger points – Evacuation concerns – Production shutdown (oil and gas)
• Excellent fire behavior modeling tools available (FlamMap, FARSITE, BehavePlus) but none appropriate for this particular task
• Fire rate of spread is available – why not turn this into time?
http://www.isciencetimes.com/articles/5392/20130612/colorado-fires-force-2-000-homes-4.htm
http://news.yahoo.com/thousands-flee-colo-wildfire-92-homes-destroyed-084723709.html
TUFA – What is it?
• Modeling technique that makes lines of equal time (isochrones) for a fire to arrive at a given location
• Works for fires located anywhere on the map
• Worst case scenario – Wind blowing toward focus
area – 97th percentile weather
scenario (adjustable) – Shortest time until fire
arrives
Methodology Overview
• Prepare input data – Fuel (LANDFIRE) – Weather – Topography
• Run FlamMap • Export Rate of Spread • Convert to time • Adjust for vectoring • Run cost distance • Visualize outputs
Fire Behavior Modeling
Fuel Weather Topography
Flame Length Rate of Spread Crown Fire
Model Builder
• Great for non-programmers!
• Visual nature makes it easy to follow
• Readily available in basic ArcMap
• Once built, the model can be run quickly
• NOT magic (see next slide)
Challenges
• Vectoring – Reconciling wind
and slope direction – Cyclic nature of
directional data
• Input data – Projection is crucial – Input data are
sometimes cranky (focus as point vs. polygon)
Limitations
• Only takes flaming front into account (not spotting/embering)
• Worst case scenario might be too extreme
• Difficult to relate model conditions to those experienced on the particular day of a fire
jalcornphoto.photoshelter.com
Case Study 1 – Future Housing Development
• Access route as focus
• Mostly one way in and out
• Negotiations with county for mitigation requirements – Second access
road? – Landscape fuel
treatment? – Treat fuel along
road?
Case Study 2
• Theme park with a large number of visitors
• Aerial tram as main access
• Shelter-in-place vs. full scale evacuation (depends on time)
Future Directions/Applications
• Other disasters? • Better handling of
slope • Address multiple
source problem • Refine fire behavior
modeling
http://www.listal.com/list/pets-steal-show
Questions or Comments?
? http://www.jimmyfungus.com/2012/11/epic-facepalm-compilation-most-epic.html