A MULTISTAGE STOCHASTIC-PROGRAMMING MODEL FOR SUSTAINABLE FOREST-LEVEL TIMBER SUPPLY UNDER RISK OF FIRE

Forest management planning models are highly developed and used extens ively, but few explicitly consider the effects of fire and other uncer tain losses which can be significant. Previous studies recommended con tradictory responses to potential fire loss. We developed forest-level timber management optimization models with deterministic average and stochastic fire loss and used them to develop insight into the impact of uncertainty on forest management planning. Our analysis indicated t hat producing a stable timber supply requires the establishment of a b uffer stock of timber. This strategy reduces the short-term harvest qu antity, but increases the expected long-term harvest quantity. Paradox ically, the timber supply is both more stable, and larger. The results are sensitive to the method of regulating timber production: profit m aximizing behavior of future producers requires a much greater buffer stock. In addition, bounds on the effect of transportation were develo ped, conditions where the mean value problem is adequate were identifi ed, and the effect of a downward sloping demand curve was investigated .