Cooperative Fish and Wildlife Research Units Program: Georgia
Education, Research and Technical Assistance for Managing Our Natural Resources

Management of ephemeral habitat conditions for endangered forest species is difficult in the face of environmental stochasticity and uncertainty about the fundamental processes of forest dynamics. Managers of southern U.S. forests who are charged with maintaining habitat for the endangered red-cockaded woodpecker must plan for an uninterrupted supply of old-growth forest stands that provide the critical breeding and foraging habitat for the bird. We constructed and optimized a dynamic model of forest growth and hardwood succession at the Piedmont National Wildlife Refuge (Georgia, USA) for the objective of maintaining a maximum stream of old-growth forest habitat over the infinite time frame. Our model accommodates stochastic disturbances and hardwood succession rates, as well as uncertainty about the structure of the model itself. We produced a decision policy that was indexed not only by current state of the forest, but also by current weight of evidence toward or against alternative forms of the model. We employed adaptive stochastic dynamic programming, which anticipates that model probabilities, as well as forest states, may change through time, with the result that the optimal decision for a given forest state can evolve over time. In light of considerable uncertainty about forest dynamics, we analyzed a set of models that spanned extremes in parameter values. Nevertheless, our analyses suggested that, under any of these models, forest silviculture practices currently employed for woodpecker recovery are suboptimal for the creation of woodpecker habitat. We endorse a fully adaptive approach to the management of endangered species habitats, in which predictive modeling, monitoring, and assessment are tightly linked.