Human encroachment into wilderness areas can influence the persistence of wildlife populations by decreasing and degrading habitat, displacement, and decreasing survival. For bighorn sheep (Ovis canadensis), some human activities are detrimental, causing both physiological stress and habitat abandonment. Between 1979 and 2000, human recreation has increased over 300% in areas occupied by desert bighorn sheep (O. c. nelsonii) in southeastern Utah. We investigated if an increase in human activity in areas used by bighorns affected their behavior. We observed 34 bighorn sheep using focal-animal sampling for >14 hrs to compare time spent grazing and scanning between areas of high and low human use. We identified group size, presence or absence of a lamb, distance to escape terrain, and human use (high versus low) as potential explanatory variables that influenced grazing and scanning times, and created an a priori list of models based on these variables. We used Akaike's Information Criterion adjusted for small sample sizes (AICc) to rank models, and used model selection to find a best approximating model (lowest AICc value) for both behaviors. Desert bighorn sheep spent less time grazing and more time scanning in high human use areas (22% grazing, 29% scanning) than in low human use areas (54% grazing, 8% scanning). Caution should be taken when considering which areas or trails should be opened during these important seasons to minimize and reduce additional stresses to bighorns caused by human activity. Bighorn sheep populations experienced significant declines after European settlement in North America. Today, the primary practice of bighorn sheep conservation is through population restoration and augmentation from remnant source populations. We conducted a 9-year telemetry study for a source population of desert bighorn sheep in Canyonlands National Park, Utah. We captured and collared 58 bighorn sheep from 2002-2009. To estimate annual and seasonal survival, we used known-fate analysis in Program MARK 4.1. We used model selection to test hypotheses for bighorn survival, including sex, age, human use, year, and month, as possible explanatory variables. There were 20 mortalities during the study. Annual survival ranged from 83% - 88% with no significant variation among any of the years. Model selection results showed that the top six models included a temporal variable (e.g. season or month), and carried 92% of the AICc weight. Population persistence for bighorn sheep can be compromised by high levels of predation, habitat fragmentation, and disease transmitted from domestic sheep. We suggest that land managers continue to maintain the separation of domestic sheep from bighorns in CNP. We also recommend that survival studies continue to ensure that future translocation projects do not occur at the expense of the source population.