Forest management under uncertainty for multiple bird population objectives
We advocate adaptive programs of decision making and monitoring for the management of forest birds when responses by populations to management, and particularly management trade-offs among populations, are uncertain. Models are necessary components of adaptive management. Under this approach, uncertainty about the behavior of a managed system is explicitly captured in a set of alternative models. The models generate testable predictions about the response of populations to management, and monitoring data provide the basis for assessing these predictions and informing future management decisions. To illustrate these principles, we examine forest management at the Piedmont National Wildlife Refuge, where management attention is focused on the recovery of the Red-cockaded Woodpecker (Picoides borealis) population. However, managers are also sensitive to the habitat needs of many non-target organisms, including Wood Thrushes (Hylocichla mustelina) and other forest interior Neotropical migratory birds. By simulating several management policies on a set of-alternative forest and bird models, we found a decision policy that maximized a composite response by woodpeckers and Wood Thrushes despite our complete uncertainty regarding system behavior. Furthermore, we used monitoring data to update our measure of belief in each alternative model following one cycle of forest management. This reduction of uncertainty translates into a reallocation of model influence on the choice of optimal decision action at the next decision opportunity.