US Geological Survey
Date of this Version
2002
Citation
Published in Wildlife Society Bulletin (2002) 30(1): 176-185
Abstract
In 1995, the United States Fish and Wildlife Service adopted a protocol for the adaptive management of waterfowl hunting regulations (AHM) to help reduce uncertainty about the magnitude of sustainable harvests. To date, the AHM process has focused principally on the midcontinent population of mallards (Anas platyrhynchos), whose dynamics are described by 4 alternative models. Collectively, these models express uncertainty (or disagreement) about whether harvest is an additive or a compensatory form of mortality and whether the reproductive process is weakly or strongly density-dependent. Each model is associated with a probability or "weight," which describes its relative ability to predict changes in population size. These Bayesian probabilities are updated annually using a comparison of population size predicted under each model with that observed by a monitoring program. The current AHM process is passively adaptive, in the sense that there is no a priori consideration of how harvest decisions might affect discrimination among models. We contrast this approach with an actively adaptive approach, in which harvest decisions are used in part to produce the learning needed to increase long-term management performance. Our investigation suggests that the passive approach is expected to perform nearly as well as an optimal actively adaptive approach, particularly considering the nature of the model set, management objectives and constraints, and current regulatory alternatives. We offer some comments about the nature of the biological hypotheses being tested and describe some of the inherent limitations on learning in the AHM process.