Fences reduce habitat for a partially migratory ungulate in the Northern Sagebrush Steppe

Abstract

Few studies have examined differential responses of partially migratory ungulates to human development or activity, where some individuals in a population migrate and others do not. Yet understanding how animals with different movement tactics respond to anthropogenic disturbance is key to sustaining global ungulate migrations. We examined seasonal resource selection of a partially migratory population of pronghorn (Antilocapra americana) in the Northern Sagebrush Steppe of Alberta, Saskatchewan, and Montana from 2003 to 2011. We developed step‐selection functions (SSF) for migrant and resident pronghorn during the summer and winter at two spatial scales (second order and third order) and then integrated SSFs across scales to estimate pronghorn responses to fences and subsequent habitat loss from these features while accounting for responses to other resource use. Both migrant and resident pronghorn showed the strongest responses to natural and anthropogenic features at the second order and weaker responses at the third order. Selection responses of migrant and residents differed the most in response to normalized difference vegetation index, topography, and anthropogenic features. Seasonally, selection for intermediate greenness was strongest in summer, whereas avoidance of roads strongly influenced winter resource selection of both tactics. Both migrant and resident pronghorn showed strong avoidance of fencing at both spatial scales during summer and winter. Model predictions with complete removal of fences from the landscape (i.e., natural conditions) predicted an increase in the area of high‐quality habitat of 16–38%. In contrast, doubling fence density on the landscape decreased the amount of high‐quality habitat by 1–11% and increased low‐quality habitat by 13–21%. Our results suggest that pronghorn winter and summer ranges can be improved by reducing the density of fences on the landscape, or mitigation measures to enhance fence crossings, to alleviate the indirect loss of habitat for this important endemic prairie species.

Multi-scale habitat assessment of pronghorn migration routes

“…In addition, we found selection for grasslands across scales during spring and at broader scales only during fall migratory periods, but surprisingly not NDVI, which was only included in spring migration models. Our results vary to some degree to those reported by [ 35 ] for our migratory pronghorn during summer and winter seasons. They found migratory pronghorn in the Northern Sagebrush Steppe to consistently select against paved and all roads across both seasons and scales of selection, while roads where only important at the broad-scale during our study [ 35 ].…”

“…Our results vary to some degree to those reported by [ 35 ] for our migratory pronghorn during summer and winter seasons. They found migratory pronghorn in the Northern Sagebrush Steppe to consistently select against paved and all roads across both seasons and scales of selection, while roads where only important at the broad-scale during our study [ 35 ]. During summer, NDVI was significant across scales for migratory pronghorn, which is only consistent with our spring migration period results [ 35 ].…”

“…We found that neither forage productivity (measured using NDVI) or snow cover were significant factors influencing selection though both have been suggested to initiate fall migration [68,72,73]. While we did not demonstrate a vegetation senescence factor in driving fall migration, [35] did report NDVI as a significant factor in the selection of winter ranges by Northern Sagebrush Steppe migratory pronghorn. Annual snow cover variation was observed across the study area, most notably during the fall period (October-November), which may be a prominent factor to observe yearly differences in migratory start and end dates, as well as overall distances moved during fall migrations [24].…”

“…First, we calculated the well densities specific to time period across the study period and second, calculated the distance of each used and available location from wells actively being drilled within a temporal window of 16 days to account for human activity during drilling and extraction [ 52 ]. We also evaluated quadratic terms for road and slope variables to account for possible non-linear responses by pronghorn [ 28 , 35 ].…”

“…During both migrations periods we predicted that pronghorn would follow spatiotemporal gradients in forage productivity (phenology) [ 31 , 32 ] along steps within the migratory pathway (third-order), while at the scale of a migratory neighborhood (second-order), we predicted pronghorn would select for native grasslands in less rugged areas to maintain energy reserves required for migration [ 25 , 33 ]. We also predicted pronghorn would avoid higher densities of roads and oil and natural gas wells during migration [ 28 , 34 , 35 ]. Lastly, we predicted seasonal variation with a stronger effect of forage phenology during spring green-up compared to fall migration [ 24 , 36 ].…”