To measure Northern Goshawk (Accipiter gentilis) home range sizes and habitat selection in naturally-fragmented forests of south central Idaho, I radio-tracked six adult males throughout their breeding seasons in 2001 and 2002. Males were tracked from the time young were 5-10 d old until 20 d post-fledging. I collected 485 temporally-independent locations, with a mean of 54 3.3 (SE) such locations per bird. Average linear error associated with triangulated locations was 130 15 m. Median minimum convex polygon home range size was 588 ha for males that successfully fledged young, and 542 ha for unsuccessful males. Median 85% cluster core area size was 98 ha for successful males, and 70 ha for unsuccessful males. Home ranges of neighboring males overlapped by an average of 19%. These home ranges were smaller than any previously reported for goshawks in North America and may have been due to a super abundance of hunting areas, such as edge areas where forest and shrub steppe habitats interfaced and goshawks perched to hunt open-country prey, especially ground squirrels.
I assessed goshawk habitat selection at two scales [perching/hunting sites within the home range (Johnsons 3rd order), and home range placement within the larger study area (Johnsons 2nd order)]. Habitat variables measured included: distance from used (obtained through radio telemetry) and available (randomly selected within home ranges) locations to the nearest seedling (< 2.5 cm dbh) stand, sapling/pole (2.6-12.9 cm dbh) stand, “small tree” (13-34 cm dbh) stand, road/trail, and camp area. Small tree habitat constituted the largest trees in the study area. I used logistic regression to identify those features that may be important in predicting individual goshawk use of a perching site. Five out of six males selected perching sites closer to “small tree” habitat and camp areas than expected. Two males selected perching locations closer to roads/trails than expected, and two males selected higher or lower elevations than expected. I examined habitat selection at the scale of home range placement within the study area using home-range sized sampling ellipses placed around 23 known historical nests, and 23 randomly selected locations within the study area. At this scale, goshawks selected for less sapling/pole habitat, and greater habitat diversity as measured by Shannon’s diversity index, similar to results from other studies.