Date of Award
Quantifying the abundance and distribution of animal populations is critical for effective wildlife research and management. Due to their cost-effectiveness, wildlife cameras have become an increasingly popular tool for estimating population densities. Previously, this technique relied on ‘capture-recapture’ models that utilized re-sightings of individually marked animals, but in recent years methods have been developed to estimate the population densities of unmarked animals. One such method is the random encounter and staying time (REST) technique, which does this by assuming that the cumulative time animals stay within the view of the camera scales linearly with the number of individuals. This allows for a density estimate without the need to determine individual identity. To evaluate the accuracy and precision of the REST method, I compared cattle (Bos taurus) density estimates based on trail-camera photos to the actual number of cattle stocked on a U.S. Forest Service (USFS) grazing allotment. Photos were collected across 96 motion-activated cameras distributed across a single grazing allotment in Spanish Fork, Utah. Based on the USFS grazing plan, the allotment operated under a rest-rotation grazing system, and therefore was divided into three pastures, only one of which held cattle at any given time in the year. Based on this plan cattle numbers also varied throughout the year according to a set schedule. For each stocking period and pasture, we generated REST-based abundance estimates, including empirical confidence bounds derived using either spatial or temporal averaging. Our results indicate very poor agreement between REST-based estimates and USFS stocking rates, where, at the allotment level, the former are typically 50-350% higher than the latter. Whether this indicates REST-based estimates are biased or inaccurate is hard to say; there is no doubt our cameras had detected cows (sometimes a lot of cows) in places and times that no cows should have been in based on USFS records. We thus have little confidence in the reliability of these records. As for precision, coefficient of variation values for our estimates ranged between 0.1 and 0.5 (depending on the number of active camera days used to calculate the estimate, and on whether densities were averaged across space or across time). This indicates that REST-based estimates are at least precise enough to be reasonably consistent across time (and to a lesser degree, space), and may hence be a valuable tool at the hand of wildlife managers.
Bonebrake, Emily, "Estimating Cattle Density Using Wildlife Cameras" (2022). Undergraduate Honors Capstone Projects. 927.
Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .
Departmental Honors Advisor