Application of hazard and risk analysis at the project level to assess ecologic impact

Abstract The application of hazard and risk analysis to specific project areas prone to uncharacteristic wildland fires is a useful way to estimate the effects of management alternatives (including no action). These project-level analyses need to be done in the context of surrounding landscape conditions. A landscape-level analysis is often at the catchment scale or larger, while project work is generally at a smaller scale, limited by practical considerations such as budget, land ownership patterns and public perception. This difference in scale requires an interpretative procedure to select an ecologically effective project alternative, and we propose a decision process involving several steps of hazard and risk analysis. The first step is to evaluate wildfire hazard and risk elements at the landscape level over longer time frames to provide insight into the factors dominating fire behavior and the most imperiled physical or ecologic domain such as vegetative succession, watershed values or human health and safety. Second, we suggest an additional spatial consideration to estimate the representative elemental scale (RES) of the fire process in the landscape. Consideration of the RES allows estimation of project-scale impacts to landscape-scale problems, while considering the hazard and risk assessment helps estimate longer-term project impacts, and possible cumulative impacts from multiple project activities. Third, we propose considerations and objective functions to be used in locating and sizing project areas, and applying treatment prescriptions to specific situations within the project area. The latter steps require fire history data and output from a fire behavior or vegetative succession computer model. We use data from the Southwest Oregon Demonstration Project [Roloff, G.J., Mealey, S.P., Clay, C., Barry, J., Yanish, K., Neuenschwander, L. A process for modeling short- and long-term risks in the Southern Oregon Cascades, submitted for publication.] to illustrate the methods proposed. Roloff et al. demonstrate a formal model incorporating these concepts.