摘要：The effectiveness of conservation efforts ultimately depends on the recognition that decision making, and the systems that it is designed to affect, are inherently dynamic and characterized by multiple sources of uncertainty. To cope with these challenges, conservation planners are increasingly turning to the tools of decision analysis, especially dynamic optimization methods. Here we provide a general framework for optimal, dynamic conservation and then explore its capacity for coping with various sources and degrees of uncertainty. In broadest terms, the dynamic optimization problem in conservation is choosing among a set of decision options at periodic intervals so as to maximize some conservation objective over the planning horizon. Planners must account for immediate objective returns, as well as the effect of current decisions on future resource conditions and, thus, on future decisions. Undermining the effectiveness of such a planning process are uncertainties concerning extant resource conditions (partial observability), the immediate consequences of decision choices (partial controllability), the outcomes of uncontrolled, environmental drivers (environmental variation), and the processes structuring resource dynamics (structural uncertainty). Where outcomes from these sources of uncertainty can be described in terms of probability distributions, a focus on maximizing the expected objective return, while taking state-specific actions, is an effective mechanism for coping with uncertainty. When such probability distributions are unavailable or deemed unreliable, a focus on maximizing robustness is likely to be the preferred approach. Here the idea is to choose an action (or state-dependent policy) that achieves at least some minimum level of performance regardless of the (uncertain) outcomes. We provide some examples of how the dynamic optimization problem can be framed for problems involving management of habitat for an imperiled species, conservation of a critically endangered population through captive breeding, control of invasive species, construction of biodiversity reserves, design of landscapes to increase habitat connectivity, and resource exploitation. Although these decision making problems and their solutions present significant challenges, we suggest that a systematic and effective approach to dynamic decision making in conservation need not be an onerous undertaking. The requirements are shared with any systematic approach to decision making—a careful consideration of values, actions, and outcomes.
