Environmental variation in mountainous regions can impose major differences in demography and physiology on animal populations that occupy a large elevation range. This variation can be both predictable and unpredictable. In the south-western Yukon, arctic ground squirrel (AGS) populations occur all the way from the forested valley bottoms to the alpine meadows and, in so doing, experience a wide range of predation risk, forage quality and exposure to weather variables. To assess how AGS integrate these factors and respond to them, we measured fecal glucocorticoid metabolites (FCM) as an index of the stress axis on four sites at different elevations. We first validated the enzyme immunoassay with a hormonal challenge protocol and a simulated predator stressor. AGS show a strong diurnal pattern in FCM levels, with peaks at mid-day and a marked increase in response to adrenocorticotropic hormone and the simulated predator and a decline in response to dexamethasone. The lag time between the challenge and its reflection in the feces was 4–12 h. Using this method in our field studies, we found that FCM levels decreased as the active season progressed; however, specific patterns differed among sites. We hypothesized that the early season peak in FCM levels followed by the general decline was due to brief, intense early season breeding, followed by the necessity of AGS to increase mass in preparation for hibernation. Although we found no clear, single explanation for the different FCM patterns among sites, we hypothesized that differences in seasonal climate and adverse weather may be major factors affecting FCM levels. The environment was markedly different between years, with 2008 being colder and wetter than 2009 and this was associated with AGS in 2008 having much higher FCM levels in general than 2009. We found that population density and visibility may also contribute. In conclusion, AGS live in a mosaic of habitats and each population is faced with a variety of environmental stressors; how they cope and respond to these stressors may not depend on a single factor but the complete aggregate of these stressors.