|Abstract or Summary
- Mule deer (Odocoileus hemionus hemionus) and Rocky Mountain elk (Cervus canadensis nelsoni; hereafter elk) populations in northeast Oregon have declined in the past 10 to 20 years. Concurrent with these declines, cougar (Puma concolor) populations have apparently increased, leading to speculation that predation by cougars may be responsible for declining ungulate populations. However, empirical data on cougar diets, kill rates, and prey selection are lacking to support this speculation. Furthermore, the common assumption that cougar populations have increased in northeast Oregon may not be well founded because cougar populations in other areas within the Pacific Northwest region have declined in recent years. My primary research objectives were to (1) estimate kill rates and prey selection by cougars in northeast Oregon, (2) document causes of mortality and estimate survival rates for cougars, (3) estimate population growth rates of cougars in northeast Oregon and simulate the effects of hypothetical lethal control efforts on the cougar population, and (4) investigate the relative influence of top-down, bottom-up, and climatic factors for limiting population growth rates of elk in northeast Oregon. Results from my research will help guide cougar and elk management in northeast Oregon and provide a framework for assessing relative effects of top-down, bottom-up, and abiotic factors on population growth rates of ungulates in this and other areas.
I implemented a 3-year study in northeast Oregon to investigate diets, kill rates, and prey selection of cougars in a multiple-prey system to better understand mechanisms by which cougars may influence ungulate populations. During my research, 25 adult cougars were captured and fitted with Global Positioning System (GPS) collars to identify kill sites. I monitored predation sequences of these cougars for 7,642 days and located the remains of 1,213 prey items killed by cougars. Cougars killed ungulates at an average rate of 1.03 per week (95% CI = 0.92 – 1.14); however, ungulate kill rates were variable and influenced by the season and demographic classification of cougars. Cougars killed ungulates 1.55 (95% CI = 1.47 – 1.66) times more frequently during summer (May-Oct) than during winter (Nov-Apr), but killed similar amounts of ungulate biomass (8.05 kg/day; 95% CI = 6.74 – 9.35) throughout the year. Cougars killed ungulates more frequently in summer because juvenile ungulates comprised most of the diet and were smaller on average than ungulate prey killed in winter. Female cougars with kittens killed more frequently (kills/day) than males or solitary females. After accounting for the additional biomass of kittens in cougar family groups, male cougars killed on average more biomass of ungulate prey per day than did females (R = 0.41, P < 0.001), and female cougars killed more biomass of prey per day as a function of the number and age of their kittens (R = 0.60, P < 0.001). Patterns of prey selection were influenced by season and demographic classification of cougars. Female cougars selected elk calves during summer and deer fawns during winter. In contrast, male cougars selected elk calves and yearling elk during summer and elk calves during winter. My results strongly supported the hypothesis that cougar predation is influenced by season, gender, and reproductive status of the cougar and these patterns in cougar predation may be generalizable among ecosystems. The observed selection for juvenile elk and deer suggested a possible mechanism by which cougars could negatively affect population growth rates of ungulates.
I investigated survival and documented causes of mortality for radio-collared cougars at 3 study areas in Oregon during 1989 – 2011. Mortality due to hunter harvest was the most common cause of death for cougars in the Catherine Creek study area and the study area combining Wenaha, Sled Springs, and Mt. Emily Wildlife Management Units (WSM study area) in northeast Oregon. In contrast, natural mortality was the most common cause of death for cougars in the Jackson Creek study area in southwest Oregon. Annual survival rates of adult males were lowest at Catherine Creek when it was legal to hunt cougars with dogs (Ŝ = 0.57), but increased following the prohibition of this hunting practice (Ŝ = 0.86). This latter survival rate was similar to those observed at Jackson Creek (Ŝ = 0.78) and WSM (Ŝ = 0.82). Regardless of whether hunting of cougars with dogs was permitted, annual survival rates of adult females were similar among study areas (Catherine Creek Ŝ = 0.86; WSM Ŝ = 0.85; Jackson Creek Ŝ = 0.85). I did not document an effect of age on cougar survival rates in the Catherine Creek study area, which I attributed to selective harvest of prime-aged, male cougars when it was legal to hunt cougars with dogs. In contrast, I observed an effect of age on annual survival in both the WSM and Jackson Creek study areas. These results indicate that sub-adult males had significantly lower survival rates than sub-adult females, but survival rates of males and females were similar by age 4 or 5 years. My results suggest that survival rates of cougars in areas where hunting cougars with dogs is illegal should be substantially higher than areas where use of dogs is legal.
I used estimates of cougar vital rates from empirical data collected in northeast Oregon to parameterize a Leslie projection matrix model to estimate deterministic and stochastic population growth rates of cougars in northeast Oregon when hunting cougars with dogs was legal (1989 - 1994) and illegal (2002 - 2011). A model cougar population in northeast Oregon that was hunted with dogs increased at a mean stochastic growth rate of 21% per year (λ[subscript s] = 1.21). Similarly, I found that a model cougar population that was subjected to hunting without dogs increased at a rate of 17% per year (λ[subscript s] = 1.17). Given that hunting cougars with dogs typically results in increased harvest and reduced survival rates of cougars, it was unexpected that the cougar population subjected to hunting with dogs was increasing at a faster rate than one that was not hunted with dogs. However, cougar populations in Oregon were subjected to low harvest rates when hunting cougars with dogs was legal and harvest was male biased. This resulted in high survival rates of female cougars and correspondingly high population growth rates.
The Oregon Cougar Management Plan allows the Oregon Department of Fish and Wildlife to administratively reduce cougar populations to benefit ungulate populations, reduce human-cougar conflicts, and limit livestock depredation. Consequently, I was interested in modeling the effects of a hypothetical lethal control effort on a local cougar population. Using empirically-derived vital rates and a deterministic Leslie matrix model, I found that the proportion of the cougar population that would need to be removed annually to achieve a 50% population reduction within 3 years was 28% assuming a closed population, and 48% assuming maximum immigration rates into the population. Using a stochastic Leslie matrix model, I also determined that the model cougar population would likely return to its pre-removal size in 6 years assuming a closed population, and 2 years assuming maximum immigration rates. These model results indicate that current management practices and harvest regulations, combined with short-term, intensive, and localized population reductions, are unlikely to negatively affect the short-term viability of cougar populations in northeast Oregon. However, at this time, it is not known if intensive lethal control efforts funded by state agencies will be cost-effective (i.e., increased sales of tags to hunt deer and elk will offset the costs of control efforts). Further research is needed to investigate the cost-effectiveness of cougar control efforts in Oregon.
I developed a Leslie matrix population model, parameterized with empirically-derived vital rates for elk in northeast Oregon, to investigate the relative influence on elk population growth rates of (1) survival and pregnancy, and (2) top-down, bottom-up, and climatic variables. I then estimated the effect of varying the strength of top-down factors on growth rates of elk populations. Growth rates of the model elk population were most sensitive to changes in adult female survival, but due to the inherent empirical variation in juvenile survival rates explained the overwhelming majority of variation in model population growth rates (r² = 0.92). Harvest of female elk had a strong negative effect on model population growth rates of elk (r² = 0.63). An index of cougar density was inversely related to population growth rates of elk in my model (r² = 0.38). A delay in mean date of birth was associated with reduced juvenile survival, but this had a minimal effect on population growth rates in my model (r² = 0.06). Climatic variables, which were used as surrogates for nutritional condition of females, had minimal effects on population growth rates. Likewise, elk density had almost no effect on population growth rates (r² = 0.002). The results of my model provided a novel finding that cougars can be a strong limiting factor on elk populations. Wildlife managers should consider the potential top-down effects of cougars and other predators as a limiting factor on elk populations.