Data from: The ecological footprint of recreation: Impacts on mountain goat habitat selection
Data files
Mar 11, 2026 version files 97.39 MB
-
Multivariate_data.csv
39.85 MB
-
README.md
6.02 KB
-
Summer_data.csv
57.54 MB
Abstract
Understanding the relationship between wildlife and their environment is important to wildlife management. Factors such as human disturbance that influence the behavior of animals have implications for management decisions and are critical to consider in evaluations of wildlife habitat selection. In Montana, there is concern regarding the status of mountain goats (Oreamnos americanus), and more information regarding mountain goat ecology is needed to guide management decisions. We used GPS locations from 14 mountain goats in the Bridger Range of southwest Montana collected from 2021-2024 to evaluate summer habitat selection and the effects of trail-based recreation. Recreation covariates representing the intensity of recreational use were developed based on trail-use counts obtained from trail counters and categorized into four categories (no-, low-, medium-, and high-recreation). First, we developed a summer habitat selection model that identified habitat selection within the Bridger Mountains. Next, we added recreation covariates to the most-supported model to evaluate the effects of trail-based recreation on mountain goat habitat selection. Finally, interactions with time of day for all covariates and day of week with recreation associated covariates were added to assess evidence for temporal (diel and weekly) variation in habitat selection. We found that mountain goats strongly selected steep and rugged terrain with low canopy cover during the summer. At the population level we found a lower relative probability of selection for areas of medium recreation intensity compared to areas of no or low recreation intensity. The relative probability of selection for high recreation intensity was similar to those of other recreation intensities, which did not provide strong evidence of avoidance, suggesting potential tolerance among some individuals. We found evidence for temporal variation in mountain goat selection of most covariates. Our results provided evidence of substantial individual variation in responses to recreation intensity. This information should help guide future management decisions related to timber management and suggest a cautious approach to the future development of recreation infrastructure given the varied responses of mountain goats to recreation and the potential for increased human-mountain goat conflict created by tolerance of human disturbance in high-quality mountain goat habitat by some individuals.
https://doi.org/10.5061/dryad.crjdfn3g2
Description of the data and file structure
GPS collars were used to track the locations of mountain goats in the Bridger Range of southwest Montana and split into separate datasets. A summer population range was estimated using the GPS locations. GPS locations were defined as use points for each dataset. Available locations were randomly sampled at a ratio of 1:5 (used:available) from the estimated summer population range.
Files and variables
File: Summer_data.csv
Description: This file contains the data used in univariate summer habitat modeling.
Variables
- DeviceID: Animal ID
- used: A binary variable for whether a point is used (1) or available (0)
- DT: This variable is the distance in meters to the nearest trail
- DistET40: This variable is the distance in meters to the nearest escape terrain, with escape terrain being defined as slopes 40 degrees or steeper
- DistET50: This variable is the distance in meters to the nearest escape terrain, with escape terrain being defined as slopes 50 degrees or steeper
- VRM100: This variable is a measure of terrain ruggedness within a 100 m radius of the point.
- VRM300: This variable is a measure of terrain ruggedness within a 300 m radius of the point.
- VRM500: This variable is a measure of terrain ruggedness within a 500 m radius of the point.
- Slope100: This variable is the mean slope within a 100 m radius of the point.
- Slope300: This variable is the mean slope within a 300 m radius of the point.
- Slope500: This variable is the mean slope within a 500 m radius of the point.
- Alpine100: This variable is the percent alpine landcover within a 100 m radius of the point.
- Alpine300: This variable is the percent alpine landcover within a 300 m radius of the point.
- Alpine500: This variable is the percent alpine landcover within a 500 m radius of the point.
- Canopy100: This variable is the percent canopy cover within a 100 m radius of the point.
- Canopy300: This variable is the percent canopy cover within a 300 m radius of the point.
- Canopy500: This variable is the percent canopy cover within a 500 m radius of the point.
- NDVI: This variable is the mean time-integrated NDVI for years 2021-2023 and represents the average forage productivity across years.
- Rad: This variable is the total incoming solar radiation for the summer season and is measured in watt-hours per meter squared.
File: Multivariate_data.csv
Description: This file contains data used in the multivariate summer habitat modeling.
Variables
- DeviceID: Animal ID
- used: A binary variable for whether a point is used (1) or available (0)
- DT: This variable is the distance in meters to the nearest trail
- DET40: This variable is the distance in meters to the nearest escape terrain, with escape terrain being defined as slopes 40 degrees or steeper
- VRM500: This variable is a measure of terrain ruggedness within a 500 m radius of the point.
- TPI: This variable is a measure of ridgelines (the largest positive values) and valleys (the lowest negative values) within a 100 m radius of the point.
- Counter: The daily average count of trail users from the nearest trail counter
- Slp100: This variable is the mean slope within a 100 m radius of the point.
- Rad: This variable is the total incoming solar radiation for the summer season and is measured in watt-hours per meter squared.
- NDVI: This variable is the mean time-integrated NDVI for years 2021-2023 and represents the average forage productivity across years.
- CC500: This variable is the percent canopy cover within a 500 m radius of the point.
- Intensity300: Recreation intensity of each location with a 300-meter buffer on each trail. This variable is a factor with four levels: 0= no recreation, 1=low recreation intensity, 2=medium recreation intensity, 3=high recreation intensity.
- Intensity500: Recreation intensity of each location with a 500-meter buffer on each trail. This variable is a factor with four levels: 0= no recreation, 1=low recreation intensity, 2=medium recreation intensity, 3=high recreation intensity.
- Bridger: A binary variable indicating if the location is within the Bridger Bowl Ski Area (1) or outside the ski area (0)
- Alp500: This variable is the percent alpine landcover within a 500 m radius of the point.
- Wkday: A binary variable indicating if the observation is from a weekday (1) or weekend (0).
- day: A binary variable indicating if the observation is during the daytime (1) or nighttime (0).
Access information
Portions of data were derived from the following sources:
-
The time-integrated NDVI covariate is derived from:
Benedict, T. D., D. Shrestha, and S. P. Boyte. 2021. “S-NPP 375-m eVIIRS Remote Sensing Phenology Metrics - across the Conterminous U.S. U.S. Geological Survey Data Release.” Raster Dataset. Phenology Metrics. U.S. Geological Survey. https://doi.org/10.5066/P9PZTNBI.
-
The canopy cover covariate is derived from:
U.S. Geological Survey and U.S. Forest Service. 2023. “NLCD CONUS 2021 Tree Canopy Cover.” Raster digital data. https://www.mrlc.gov/data.
-
The distance to trail and recreation intensity covariates are derived from:
USDA Forest Service. 2019. “TrailNFS_Publish.” Shapefile. Vector digital data. https://data.fs.usda.gov/geodata/edw/datasets.php?xmlKeyword=trailnfs.
-
The alpine covariate is derived from:
LANDFIRE: LANDFIRE Existing Vegetation Type layer. (2024, May - last update). U.S. Department of the Interior, Geological Survey, and U.S. Department of Agriculture. [Online]. Available: https://landfire.gov/viewer/ [2024-09-26].
These data consist of used and available points for mountain goats and the associated covariate values for each point. These data were processed using the methods outlined in the supplementary information of Macdonald et al.