Thursday, August 7, 2008 - 4:00 PM

COS 96-8: A probabilistic wildfire risk model for Los Alamos National Laboratory

Randy G. Balice, Scot D. Johnson, Kathryn D. Bennett, and Todd L. Graves. Los Alamos National Laboratory

Background/Question/Methods

Los Alamos National Laboratory (LANL) is located in the eastern Jemez Mountains, New Mexico, a region that is characterized by high frequencies of wildfires.  However, the risks presented by wildfire are difficult to quantify.  The objective of this project was to develop a spatial-temporal modeling system that expresses wildfire risks in terms of probabilities.  This facilitates the comparison and prioritization of risks for the Los Alamos region, and provides a basis for management decision-making.  We accomplished this objective by developing probability models and spatial algorithms of lightning-caused fires, using data representing lightning, weather, fuels and other parameters related to wildfire hazards.  The final output of the model was in the form of a probability value for each pixel in a regional land cover map.  These risk values were calculated as the joint probability of 1) the occurrence of a suitable lightning flash, 2) the suitability of the ambient fuel moisture conditions for a lightning flash to result in an ignition, and 3) the potential for a lightning flash to be accompanied by the absence of ignition-extinguishing rainfall.  The final risk values may be interpreted as the potential for a lightning ignition to occur and to smolder in the ground fuels for up to two days. 

Results/Conclusions

The results of these analyses indicate that the wildfire risk from lightning strikes before April 10 is very low.  The risks are highest from June 12 to July 4.  The risks were low to intermediate during the other time periods. The highest risk levels were approximately 0.355, and these levels occurred in the mountainous areas to the west of LANL.  A composite risk over the entire wildfire season was also calculated and mapped.  The highest composite risk was 0.49, and these greater risks tended to occur in the higher elevations to the west of LANL.  With regard to the LANL and its immediate surroundings, the greatest risks were observed in the western and southwestern portions of the Lab and in the adjacent mountainous areas.