Modeling Fire Behavior in the East S.F. Bay with Flam Map and ArcGIS

Modeling Fire Behavior in the East S.F. Bay with Flam Map and ArcGIS Image: Business Insider Introduction Wild fires are a normal and sometimes necessary ecological process in a many parts of the world including California. However, wildfires are a problem in California as they threaten human life and property. SCIENCE CONNECTIONS WESTERN WILDFIRES & CLIMATE CHANGE Wildfires are increasing and wildfire season is getting longer in the Western U.S. Average number of large wildfires per year bigger than 1,000 acres California’s population is expected to increase by approximately 12.8 million people from 2014 to 2060 (D.O.F.). Urban areas will absorb a portion of this increase, but development will also continue to spread into wilderness areas. The homes in this Wildland-Urban Interface are particularly vulnerable to wildfires. Compounding this problem, wildfire frequency has been increasing due to the effects of climate change such as increasing temperatures and drought. In fact, the US Forest Service went from spending 16% of their budget on wildfires in 1995 to spending 50% in 2015 (Forest Service). 1980-1989 1990-1999 2000-2012 ~140 ~160 ~250 Average length of wildfire season Early 1970s: 5 months Today: 7+ months UCSUSA.org Climate change is driving up temperatures and increasing wildfire risk. TEMPERATURES ARE RISING Average annual temperatures in the Western U.S. have increased 1.9° F since 1970. SNOW MELTS SOONER Winter snowpack melts up to 4 weeks earlier than in previous decades. FORESTS ARE DRIER, LONGER Conditions are primed for wildfires to ignite and spread. Wildfires are projected to burn more land as temperatures continue to rise. Projected increase in annual burn area with an additional 1.8° F rise in temperature 0% 200% 400% 650% By mid-century, temperatures in the Western U.S. are expected to increase even more (2.5°–6.5° F) due to heat-trapping emissions from human activity. The choices we make today will determine how much temperatures increase this century, how long and damaging wildfire seasons become, and how prepared communities are for the growing risks of wildfires. © Union of Concerned Scientists 2013; www.ucsusa.org/westernwildfires Study Area The East Bay Hills area is located to the East of the San Francisco Bay, CA. The region has warm, dry Summers characteristic of the Mediterranean climate in California. The hills have a peak elevation of 1905 feet, with slopes that can aid the spread of a fire. The area is also characterized by an excess of unmanaged, flammable vegetation such as pine and eucalyptus trees, and overgrown brush. All of these factors combine to make the wildland urban interface of the East Bay Hills an area particularly prone to fire. The Tunnel Fire that burned here in 1991 burned 1520 acres, killed 25 people, destroyed or damaged 3469 homes, and cost an estimated 1.5 billion dollars (SF Museum). 1991 Tunnel Fire - Berkeley / Oakland Project Purpose Planning and land use policies can address this issue and help prevent unnecessary losses of life and property through the development of vegetation management plans. To inform these planning efforts, fire characteristics and movement are modeled within a GIS. Understanding the movement of wildfires in WUI areas in particular requires large scale data in order to capture the variability of land cover. The purpose of this project is to help identify priority areas of interest for vegetation management application within the study area in the East Bay Hills. Methods The methods for this project were to use ArcGIS and Flam Map fire modeling software. GIS data of certain environmental characteristics of the study area was developed and then combined into a fuel model. This was input into Flam Map, which output metrics of fire rate of spread, flame length, and canopy fire for the study area. The GIS data used in creating the fuel model was developed by the students in the LA C188 course. I collaborated with Professor Radke in developing the lab assignment. I developed a tutorial video for the assignment as well. Each student was assigned a unique site area for which they digitized every feature. Feature classes were created for tree canopy, shrubs, open green space, buildings, pavement and pools. Various attributes were recorded such as tree heights and species, and building material types such as wood shingles and siding. The data from the student submissions was then assembled and processed into a fuel model raster and a percent canopy cover raster, both incorporating the unique site characteristics and variability of land cover. Fuel Model as Vectors on TIN (also previous page image) The fuel model raster contains a value for each pixel that represents the unique fire behavior characteristics associated with that land cover type. The values correspond to the 13 Anderson fire behavior fuel model codes. (See Chart) A 3ft x 3ft digital elevation model was acquired, and slope and aspect rasters were derived from it. Elevation, Slope, Aspect, Canopy Cover, and Fuel Model rasters were converted to ASCII file format and combined into a Landscape (.LCP) file for input into the fire model. The landscape file, along with fuel moisture and wind speed variables derived from the conditions of the 1991 Tunnel Fire, were combined as inputs to the Flam Map Model. Slope Model with Houses Elevation Model Conceptual Model A 3ft x 3ft digital elevation model was acquired, and slope and aspect rasters were derived from it. Elevation, Slope, Aspect, Canopy Cover, and Fuel Model rasters were converted to ASCII file format and combined into a Landscape (.LCP) file for input into the fire model. The landscape file, along with fuel moisture and wind speed variables derived from the conditions of the 1991 Tunnel Fire, were combined as inputs to the Flam Map Model. Conclusions Units are chains per hour, commonly used by foresters. One chain is equal to 66 feet. Conclusions Flame length has a strong positive correlation with fire intensity. This metric is important in determining the difficulty and consequent method of suppression at the fire front. Conclusions The Crown Fire output indicates whether each cell meets the conditions required for fire to reach and spread at the tree canopy level. Value 1 = Surface fire only with no ability to reach the canopy. Value 2 = Conditions permit flames to reach canopy, but not spread to other tree canopies. Value 3 = Conditions permit flames to reach canopy AND spread to other trees at the canopy.