Sarah R. B. King

Understanding factors driving resource selection and habitat use of different species is an important component of management and conservation. Feral horses ( Equus caballus ) are free ranging across various vegetation types in the western United States, yet few studies have quantified their resource selection and seasonal use. We conducted a study to determine effects of vegetation community, distance to water, and topographic variables on seasonal resource selection in 2 feral horse populations in Great Basin sagebrush ( Artemisia spp.) ecosystems of west‐central Utah, USA: Conger Herd Management Area (HMA) and Frisco HMA. We deployed global positioning system (GPS) radio‐collars on 38 female horses and GPS‐transmitters braided and glued into the tail hair of 14 males, collecting locations every 2 hours for 1–4 years between 2016 and 2020. We calculated home range size and core use area of social groups (harems) and bachelor males using auto‐correlated kernel density estimators for each biologically defined season (breeding, fall, and winter) per study year. We examined seasonal home range size and overlap of harem groups and bachelor males and compared movement speed of bachelors and harems among seasons. We determined seasonal resource selection in a use‐availability framework using resource selection functions. We hypothesized that horses would select for areas of high herbaceous vegetation, that water would be a key variable in resource selection models like other equids, and home range size in winter would be largest because horses can eat snow for hydration and could therefore roam farther from surface water. Mean annual home range size was 103.12 ± 37.38 km 2 (SD) for Conger harems and 117.47 ± 32.75 km 2 for Frisco harems. At Conger there was no difference in home range size between harem groups and bachelor males, but home range size was smaller in winter than other seasons, whereas winter home range size at Frisco was larger than other seasons. Bachelor males moved at higher speeds than harems during all seasons, and harem groups from both populations had lower movement speeds in winter. Harem groups had distinct winter ranges with little overlap on breeding season ranges. In both populations, all horses selected for herbaceous vegetation types and avoided forest relative to shrubland throughout the year. Harems at Frisco were consistently located closer to water sources, whereas selection for water sources by Conger harems varied seasonally, with winter having the lowest selection. Harem groups at Conger had an average of 10.6% of their home ranges outside the HMA boundary and Frisco harems had up to 66.8% outside, likely because of the horseshoe shape of Frisco HMA in which shrub meadows (foraging areas) comprise the horseshoe center, which is outside the HMA. Our results highlight the importance of water sources, which were a key predictor of horse movement patterns in our study. We emphasize the utility of telemetry devices to understand resource selection of feral horses at a fine scale, enabling management to be more targeted and facilitate planning.