Is habitat constraining bighorn sheep distribution and restoration: a case study in the Greater Yellowstone Ecosystem

Abstract

Rocky Mountain Bighorn Sheep (Ovis canadensis) restoration continues to be a challenge throughout western North America despite nearly a century of efforts dedicated to the species' recovery. A persistent problem for restoration is populations failing to expand into surrounding areas of habitat even during years of population growth. While populations can be constrained by several environmental factors and behavioral tendencies, we contend habitat availability is not the primary limiting factor. This study incorporated GPS data from bighorn sheep within the Taylor-Hilgard population in the Madison Mountain Range, located in the northwestern extent of the Greater Yellowstone Ecosystem (GYE), to develop summer and winter resource selection function (RSF) habitat models. The objective of this study was to evaluate a hypothesis that habitat was not the primary factor limiting distributions of bighorn sheep within the Madison Range by developing biologically-plausible RSF models and using covariates expected to influence selection. Multiple functional forms and spatial grains for covariates were considered and sets of summer and winter resource selection models compared using AIC subscript c. Results indicated that bighorn sheep resource selection was grain dependent, with bighorn sheep generally selecting covariates at the larger 500 m and 1,000 m spatial grains. Summer selection was characterized by rugged terrain, steep slopes, reduced canopy cover, southwestern aspects and ridgelines. Winter selection was characterized by low elevations, southwestern aspects, steep slopes, reduced canopy cover, ridgelines, high summer NDVI amplitude, and areas close to steep terrain (slopes > or = 45°). Predicted winter habitat occurred in a non-contiguous distribution primarily along low-elevation, southwest-facing aspects within the Madison Valley, and predicted summer habitat was concentrated along high elevation ridgelines. Model results were successfully validated using independent GPS data. Potential abundance for the Madison Range was estimated by linking the winter RSF to population estimates for the Taylor Hilgard and results suggested that the Range may be capable of supporting 2 to 4 times the number of bighorn sheep currently estimated. Study results supported the hypothesis that habitat was not the primary factor limiting extant bighorn sheep populations, suggesting that broader distributions within the Range are possible if novel restoration strategies are considered.