Viability analysis of endangered species: a decision-theoretic perspective

The concept of population viability is probabilistic in nature, usually being expressed through such indicators as ‘expected extinction time’, or ‘probability of survival for 1000 years’. Such one-dimensional indicators might be adequate to characterize empirically-derived extinction risk in steady-state circumstances, such as in a long-established and undisturbed reserve. But they do not serve as well for evaluating viability in dynamically changing environments, nor for appraising the circumstantial evidence of risk provided by computer simulation of a stochastic population model. To characterize risk adequately in these circumstances requires describing a qualitative pattern of risk, sorting out short-term effects, due to initial population size and environmental state, from longer-term effects related to the character and quality of habitat. Furthermore, the relevant time-scales for the analysis depend upon the processes of change present in the habitat, including persistent effects from deliberate or inadvertent anthropogenic habitat manipulation. In the present study, alternative extinction risk indices and sustainability profiles are suggested, appropriate for formulating a risk management strategy in a dynamically changing environment. The concepts presented are applied to the problem of viability assessment of the endangered northern spotted owl, in the harvested temperate rainforest landscape of the U.S. Pacific Northwest.