Long‐Term Evaluation of Cougar Density and Application of Risk Analysis for Harvest Management

Abstract

Estimates of cougar (Puma concolor) density are among the least available of any big game species in North America because of monetary and logistical challenges. Thus, wildlife managers identify cougar density estimates as a high priority need for population estimation, developing harvest guidelines, and evaluating management objectives. Cougar densities range from <1 to almost 7 cougars/100 km2; however, the magnitude of spatial and temporal variation associated with these estimates is difficult to assess because this range of densities could potentially be reported for any given population using different demographic, temporal, durational, and analytical approaches. We used long‐term global positioning system (GPS) data from collared cougars across 5 diverse study areas in Washington, USA, as the basis for calculating multiple annual independent‐aged (≥18 months) cougar densities, using consistent methods, and conducted a meta‐analysis to assist with statewide harvest guidelines. To generate specific harvest guidelines for unobserved populations at the management unit scale, we employed a Bayesian decision‐theoretic approach that minimizes statistical risk of failing to achieve a defined harvest rate. For the 16‐year field effort, we calculated 24 annual densities for independent‐aged cougars. Average annual densities ranged from 1.55 ± 0.44 (SD) cougars/100 km2 (n = 5 years) to 2.79 ± 0.35 cougars/100 km2 (n = 5 years) among the 5 study areas. Explicit delineation of the cougar population demonstrated that contribution to density can vary considerably by sex and age class. Application of a 12–16% harvest rate within the risk analysis framework yielded a potential annual harvest of 249 cougars over 91,000 km2 of cougar habitat in Washington. Given the importance of density for establishing harvest guidelines, and the degree of uncertainty in projecting derived densities to future years and unstudied management units, our approach may lessen the ambiguity of extrapolations and increase the longevity of research results. Our risk analysis can be used for a diverse array of species and management objectives and be incorporated into an adaptive management framework for minimizing management risk. Our recommendations can improve standardization in reporting and interpretation of cougar density comparisons and bring clarity to the sources of variability observed in cougar populations. © 2021 The Wildlife Society.

Untitled

“…In addition, we estimated a conservative percentage of the eastern Sierra lion population removed annually by dividing the number of independent lions removed by the effective population size (N e ) (i.e., 79, from Gustafson et al 2019). To evaluate if removal rates exceed sustainable harvest guidelines, we compared the percentage of the eastern Sierra lion population removed annually to a guideline of 14% that has been used for sustainable harvest management in Washington (Wielgus et al 2013;Beausoleil et al 2021) but also thought to be widely applicable throughout the western U.S. (Beausoleil et al 2013).…”

“…Such rapid recolonization of vacant lion habitat suggests long-term negative impacts to the number and distribution of lions in the eastern Sierra lion population from removals to protect Sierra bighorn from predation are unlikely. This observation is further supported by our estimates that (1) the majority (> 80%) of the suitable habitat for the eastern Sierra lion population does not overlap with Sierra bighorn habitat, indicating that a substantial fraction of the lion population is not subject to removal because they do not interact with Sierra bighorn, and (2) the maximum level of lion removals to date (n = 5/yr) equates to just 6.3% of the eastern Sierra lion population (based on an estimated N e of 79 [Gustafson et al 2019]), less than half of the 14% sustainable harvest threshold needed for population stability advocated by Beausoleil et al (2013), Wielgus et al (2013), and Beausoleil et al (2021. Because we used an estimate of N e to represent the abundance of the lion population, and N e is less than numerical abundance (Frankham 1995), our estimated lion removal rates should be considered quite conservative.…”

A policymaker perspective on use of a population density estimate to manage cougar harvest risk (Beausoleil et al. 2021)

“…A policymaker perspective on use of a population density estimate to manage cougar harvest risk (Beausoleil et al 2021) Beausoleil et al (2021) recently published a dataset analysis of cougar (Puma concolor) population density estimates in Washington, USA, and touted its utility for harvest management. As a Washington fish and wildlife commissioner, I find the article's thesis that using a statistical global density calculation to set cougar abundance estimates for determining harvest allowances while using the same calculation to assert risk of over-or underharvest is insufficient.…”

“…Statewide harvest is monitored by sex and age records from required pelt sealing and tooth age analysis, which contributes little information about cougar abundance. The decision-theoretic approach of Beausoleil et al (2021) leads to the nonsensical conclusion that there is more harvest mortality (and sometimes more cougar-human conflict) because cougar abundance was over-estimated.…”

“…Using a cougar density range as the risk correlate, the Beausoleil et al (2021) model reported overharvest to be a higher probability risk than underharvest. The Washington approach to cougar seasons responds to the possibility of overharvest; however, the state has no provision for dealing with the possibility of a larger than estimated, under-hunted cougar population in a management unit that, according to model assumptions, would be increasing.…”