Applicability and limitations of sensitivity analyses for wildlife management

Abstract

Abstract Sensitivity analyses that assess the impact of changing vital rates on population growth have been widely used to guide conservation. If implemented with caution, they can provide guidance as to which management actions will optimize conservation outcomes. In this review, we first focus on the commonly used proportional sensitivity and elasticity analyses that change each vital rate by equal proportions, to assess their importance for wildlife management. These types of analyses also feature potential pitfalls and limitations, including (1) Each vital rate is usually on a different scale. Without appropriate scaling this can result in a flawed evaluation of the importance of vital rates. (2) Vital rates rarely change at equal proportions in nature. This can bring about misguided management recommendations on the basis of vital rate changes that are unrealistic. (3) Proportional sensitivity analyses often do not reflect the feasibility and effectiveness of altering particular demographic parameters. Consequently, relying solely on proportional sensitivities or elasticities can lead to flawed evaluation of the importance of vital rates and thus prioritization of management options that are unrealistic or ineffective. We outline alternative approaches, which involve assessing the impact of threats, the relative demography of stable and declining populations, the effect of observable variation of vital rates on population viability, and the potential effects of feasible management scenarios. Synthesis and applications. Sensitivity analyses are useful tools to guide wildlife management. If implemented and interpreted with care, sensitivity analyses can identify key demographic parameters and threats to population viability. However, their usefulness is limited, when applied without careful evaluation as to whether the perturbations evaluated are realistic, feasible and meet the need of wildlife managers. We caution against the over‐reliance on proportional sensitivity and elasticity analyses and point to alternative approaches, including life‐stage simulation analysis, vital rate sensitivity analysis or manual perturbations.