THE RESEARCH

Background

Escalating cougar numbers in some areas and their potentially negative effect on small populations of ungulate prey in multi-prey environments create an urgent need for better predictive models of cougar prey selection. Why do cougars sometimes unexpectedly prefer an ungulate species that is not the most readily available, and why do they significantly depress small populations of secondary prey (such as bighorn sheep and woodland caribou) in some instances but not in others. Current functional response models, which explain variation in species-specific kill rates in terms of prey density, do not account for these observations, leaving managers without the tools they need to decide how or when to intervene to protect ungulate populations.

To make matters worse, managing predator-prey systems is particularly complex precisely when intervention to protect prey is an issue. Cougars are an important wilderness icon and big game species in their own right, and biologists increasingly emphasize the importance of predation by large carnivores in structuring ecosystems.  Consequently, great care must be taken in managing these low-density, wide-ranging predators, even where ungulate populations are at risk. How can we be sure to retain viable cougar populations while simultaneously limiting the drastic effect they have on certain populations of their prey?

To answer the relevant theoretical questions, and to build predictive models that can facilitate effective management, we propose to bring many of the variables most likely to influence cougar predation patterns together in a single analytical exercise. These include: season, alternate  
prey densities, the life-history characteristics of individual cougars, prey vulnerability, and the hypothesized development of different prey niches between cougars and wolves. We will pay particular attention to landscape variables, testing competing hypotheses about the effects of anthropogenic deforestation on cougars and their prey.

Our opportunity to sort out the relative influence of such a broad range of variables is unprecedented because of the combination of new technology and extensive background research in what is rapidly becoming one of the best-understood large mammal predator-prey systems anywhere: the Rocky-Clearwater forest reserve in west central Alberta. Moreover, as the northernmost study of cougars to date, this project will add significantly to our knowledge of this magnificent predator throughout its range, with particular benefit to effective management in this part of Alberta.

Methods

We will attempt to obtain between 45 and 60 “cougar years” of data. We are targeting a three-year project (November 2005-November 2008) based on the capture (and in some cases re-capture) of 15-20 cougars each year. However, capture and data-gathering rates are inherently unpredictable for a variety of reasons, including the vagaries of weather during the crucial snow-dependant capture season, the variation of funding from year to year, and the unforeseen death of collared cougars. We are thus prepared to extend data gathering into a fourth year (2008-09) if necessary.

Each cougar will be fitted with either a global position system (GPS) collar (Lotek model 4400) programmed to fix a location every 3 hours and capable of remote download (14 collars) or a VHF radio-collar (7 collars).

Specific attempts will be made to capture
cougars  that reside in areas where smaller populations of alternate ungulate prey (e.g., sheep or elk) are present. Only resident cougars will be collared; young cougars (i.e., kittens and juveniles) will not be collared because they are expected to disperse, possibly leaving the study area entirely. Age, sex, weight, condition, and reproductive status (if female) will be recorded at capture. Individual animals will be monitored for up to a year (GPS collars) or more (in the case of VHF collars or re-capture). Cougar capture efforts will be dispersed throughout the study area to encompass the variation in landscape features and prey distributions that will be required to meet our modeling objectives.

Securing data on species-specific kill rates for individual cougars will be facilitated by the GPS radiocollars. This satellite-based technology will allow us to return to predation sites (identified by GPS location clusters) in a timely manner without having to continuously follow individual animals, allowing us to avoid the small sample sizes and high labour costs associated with snow tracking. This high-tech method also makes it possible to assess kill rates year-round, remedying the paucity of kill rate information for cougars in summer that has plagued previous studies of cougar predation.

The study site’s rich Geographic Information System (GIS) provides the landscape variables needed for the modeling exercise and for testing competing hypotheses about the impact of industrial habitat fragmentation. GIS techniques also will be used to model predation risk as it relates to landscape features and to structure the spatial relationships between predator and prey by estimating Resource Selection Functions (RSFs) for both. We will correct for GPS bias and adjust for autocorrelation of location fixes in habitat selection models because of  the potential for these two factors to distort selection coefficients. We will also estimate the
density of prey availability and the  relative abundance of prey in each radiocollared cougar’s home range, since each of these are almost certain to affect kill rates, and are thus fundamental to the estimation of a multi-species functional response. Indices for the relative density of prey at the home-range scale will be constructed from a GIS layer of prey availability for the entire study site that we will create based on snow tracking and pellet count surveys in cells drawn randomly from several landscape strata. Prey RSFs will be used to cross-check this method.

Cougar-predation and habitat-use models will be compared with similar wolf data as part of a collaborative effort with other researchers, shedding light on the question of predator prey niches, providing additional information on the cumulative effects that multiple predators have on populations of their prey, and allowing for comparisons of per-capita killing rates to evaluate the relative potential of each species to impact populations of prey in west central Alberta.