GPS collar data and social-ecological feature data for examining the movement of coyotes in Los Angeles, California (2019-2021)
Data files
Feb 05, 2025 version files 504.10 MB
-
buildingdensity.tif
59 MB
-
Coyote_GPS_RSF.csv
4.61 MB
-
Coyote_GPS_SSF.csv
4.91 MB
-
disttocemeteries.tif
31.14 MB
-
disttochannels.tif
395.41 KB
-
disttogolfcourses.tif
395.41 KB
-
disttolakes.tif
395.41 KB
-
disttoparks.tif
526.51 KB
-
disttorail.tif
395.41 KB
-
disttostreamsrivers.tif
395.41 KB
-
MedInc.tif
141.61 MB
-
NDVI.tif
69.10 MB
-
nlcddeveloped.tif
15.70 MB
-
pollburd.tif
62.27 MB
-
popdens.tif
84.90 MB
-
README.md
6.37 KB
-
roaddens.tif
28.33 MB
-
StudySiteLA.zip
6.96 KB
Abstract
How societal, ecological, and infrastructural attributes interact to influence wildlife movement is uncertain. We explored whether neighborhood socioeconomic status and environmental quality were associated with coyote (Canis latrans) movement patterns in Los Angeles, California, and assessed the performance of integrated social-ecological movement models. Herein are 1) the raw GPS data for 20 coyotes collared between 2019-2021, and 2) the social-ecological feature rasters used for data analyses.
README: GPS collar data and social-ecological feature data for examining the movement of coyotes in Los Angeles, California (2019-2021)
https://doi.org/10.5061/dryad.15dv41p5j
Description of the data and file structure
These data were used in an analysis of the influence of urban social-ecological features on coyote movement in the Los Angeles region of California. GPS collar data were initially collected between 2019-2021 as part of a different project, but were not used for that project and were instead used for this analysis. Raster data were created using various sources (see data description) for this project.
Files and variables
File: disttochannels.tif
Description: Distance to flood channels, derived using ArcGIS Pro v. 3.1.1. Original source: County of Los Angeles [Public Domain].
File: disttogolfcourses.tif
Description: Distance to golf courses, created using ArcGIS Pro v. 3.1.1.
File: disttoparks.tif
Description: Distance to state, regional, and local parks, derived using ArcGIS Pro v. 3.1.1. Original sources: County of Los Angeles [Public Domain].
File: disttorail.tif
Description: Distance to railways, derived using ArcGIS Pro v. 3.1.1. Original source: California Rail Network [Public Domain].
File: disttolakes.tif
Description: Distance to lakes, derived using ArcGIS Pro v. 3.1.1. Original source: California State Geoportal [Public Domain].
File: disttostreamsrivers.tif
Description: Distance to streams and rivers, derived using ArcGIS Pro v. 3.1.1. Original source: California Department of Fish and Wildlife [Public Domain].
File: StudySiteLA.zip
Description: Shapefile of the study area, created in ArcGIS Pro v. 3.1.1.
File: Coyote_GPS_RSF.csv
Description: GPS dataset trimmed to 2 hour fix rates, used in resource selection functions.
Variables
- GpsDescription: An indicator of individual collar ID.
- GPSTime: When the location was taken.
- Latitude: Latitude of GPS point.
- Longitude: Longitude of GPS point.
File: Coyote_GPS_SSF.csv
Description: GPS dataset used to calculate movement metrics and step selection functions.
Variables
- GpsDescription: An indicator of individual collar ID.
- GPSTime: When the location was taken.
- Latitude: Latitude of GPS point.
- Longitude: Longitude of GPS point.
File: nlcddeveloped.tif
Description: Development intensity, reclassified using ArcGIS Pro v. 3.1.1. Original source: National Land Cover Database. Reclassification scheme: 0=no data, 1=undeveloped land cover classes, 2=developed: open space, 3=developed: low intensity, 4=developed: med intensity, and 5=developed: high intensity.
File: disttocemeteries.tif
Description: Distance to cemeteries, derived using ArcGIS Pro v. 3.1.1. Original source: City of Los Angeles [Public Domain].
File: roaddens.tif
Description: Road density, derived using ArcGIS Pro v. 3.1.1. Original source: data.gov [Public Domain].
File: buildingdensity.tif
Description: Building density, derived using ArcGIS Pro v. 3.1.1. Original source: Dao, V. (2020) California building footprints [Dataset] [Public Domain]. Dryad. https://doi.org/10.7280/D16387.
File: pollburd.tif
Description: Pollution burden percentile, derived using ArcGIS Pro v. 3.1.1. Original source: Cal Enviro Screen 4.0 [Public Domain].
File: NDVI.tif
Description: Normalized difference vegetation index for Spring 2021, derived using ArcGIS Pro v. 3.1.1. Original source: Landsat 8 (USGS) [Public Domain].
File: popdens.tif
Description: Population density, derived using ArcGIS Pro v. 3.1.1. Original source: census.gov [Public Domain].
File: MedInc.tif
Description: Median income, derived using ArcGIS Pro v. 3.1.1. Original sources: County of Los Angeles [Public Domain], Southern California Association of Governments [Public Domain].
Code/software
Raster data were created or reclassified using ArcGIS Pro v.3.1.1 (ESRI, 2023). Relevant spatial covariates were summarized at the home range level for each coyote using ArcGIS Pro v.3.1.1 (ESRI, 2023). All other statistical analyses were conducted in R v.4.3.2 (R Core Team 2023).
Rmd file included herein provides code for conducting the following analyses: 1) summarizing and describing differences between movement characteristics among subset groups of coyotes (such as coyotes with more vs. less anthropogenically burdened home ranges), 2) resource selection functions (for all coyotes and subset groups), and 3) step selection functions (for all coyotes).
Primary packages that were used in these analyses include: 'adehabitatHR', 'raster', 'car', 'lme4', 'amt', 'survival', 'tidyr', 'dplyr', 'sf', 'sp', 'move', 'tidyverse', 'lubridate', 'purrr', 'MuMIN', 'stats', 'ggplot2', and 'geosphere'.
Access information
Raster datasets were created using the following sources:
- National Land Cover Database (NLCD): https://www.usgs.gov/centers/eros/science/national-land-cover-database; license: Public Domain
- United States Census Bureau: Census.gov; license: Public Domain
- Data.gov; license: Public Domain
- California State Geoportal: https://gis.data.ca.gov; license: Public Domain
- California Rail Network: https://gisdata-caltrans.opendata.arcgis.com/datasets/2ac93358aca84aa7b547b29a42d5ff52_0/about; license: Public Domain
- California Department of Fish and Wildlife: https://gis.data.ca.gov/datasets/CDFW::california-streams/about; license: Public Domain
- Dao, V. (2020) California building footprints [Dataset]. Dryad. https://doi.org/10.7280/D16387; License: Public Domain
- County of Los Angeles: https://egis-lacounty.hub.arcgis.com/; License: Public Domain
- City of Los Angeles: https://geohub.lacity.org/; License: Public Domain
- Southern California Association of Governments; License: Public Domain
Methods
GPS collar data: Beginning in October 2019, 20 coyotes were outfitted with GPS collars (Ecotone, solar powered, GPS/GSM/UHF), with collars remaining active between 1-23 months. Six females and 14 males were collared.
Raster data: All geospatial covariates were rasterized and converted to 30m2 spatial resolution using ArcGIS Pro v 3.1.1 (ESRI 2023).
Ecological covariates: (1) normalized difference vegetation index (NDVI) from spring 2021 (Landsat 8); (2) distance to rivers and streams (CDFW, 2020); (3) distance to lakes (California State Geoportal, 2021); and (4) distance to green spaces, including (a) Los Angeles and San Bernardino county parks (County of Los Angeles), (b) golf courses, and (c) cemeteries (City of Los Angeles, 2023). Distance to green spaces was considered separately from NDVI since in arid regions green spaces will not always have a notable vegetation greenness signature.
Linear infrastructure covariates: (1) road density (data.gov); (2) distance to storm and flood channels and drains (County of Los Angeles, 2023); and (3) distance to railways (California Rail Network, 2022).
Societal covariates: (1) human population density (Census.gov); (2) building density (Dao, 2020); (3) development intensity (NLCD, 2022); (4) median income (County of Los Angeles, 2023; Southern California Association of Governments); and (5) pollution burden percentile (Cal Enviro Screen 4.0). Development intensity was reclassified as 0=no data, 1=undeveloped land cover classes, 2=developed: open space, 3=developed: low intensity, 4=developed: med intensity, and 5=developed: high intensity. Cal Enviro Screen provides a pollution burden index that is calculated from 13 metrics related to drinking water characteristics, groundwater quality, air quality, soil pollutants, and hazardous waste.