Abstract: |
Forest fire managers are faced with the difficult task of choosing, from sets of suitable fuel treatments for each stand, the combination of treatments across a landscape that provides the greatest protection to the values at risk given a limited fuels treatment budget. Because wildfires are highly random events and protection is uncertain, the problem combines challenging chance constraint or chance minimization ideas with challenging scales for spatial optimization. A major complication is that the probability distribution of landscape-wide fire costs will vary with treatment patterns. We describe an approximate optimization approach to solving this problem for real-scale landscapes and we report results for some hypothetical test cases. We will begin testing the method on the Kootenai and the Idaho Panhandle National Forests this spring and possibly in Rocky Mountain National Park after that. |
