| Summary: | Degree: Master of Science Abstract: Reliable population density estimates are critical for ecological research and species management but can be difficult to obtain. Sampling methods like noninvasive genetic sampling and remote camera traps, combined with appropriate statistical models, provide opportunities to estimate density from a variety of approaches. However, it is unknown if these methods result in similar density estimates and precision of estimates. I applied and compared methods for estimating Canada lynx (Lynx canadensis) density for a cyclic population in southwestern Yukon Territory. Canada lynx are a species of ecological, economic, and conservation interest, but few studies have estimated density of lynx and even fewer have used contemporary methods. I collected lynx data using hair snares, camera traps, track transect counts, and GPS collars, then applied and compared density estimation methods across data types. Estimation methods included linearly-scaled count methods, spatial mark-recapture, spatial mark-resight, and a cumulative time method. I calculated six estimates and found five-fold variation in point estimates and two-fold variation in precision, despite closely following the methods described in current literature and making every effort to meet model assumptions. My results indicate that a single approach to wildlife density estimation is likely insufficient, and that density estimation requires careful consideration of methodological assumptions and sources of error. Specialization: Ecology
|
|---|