https://doi.org/10.5061/dryad.cjsxksndn

Environmental and anthropogenic variables associated with bears' and wolves' Global Positioning System locations in the central Apennines, Italy (2005-2010).

Description of the data and file structure

The dataset refer to environmental and anthropogenic variables recorded in a GIS environment on points (i.e., locations) used ("Pres"=1) and available ("Pres=0) to sympatric adult bears (Ursus arctos marsicanus, n=16) and wolves (Canis lupus, n=7) living in the Abruzzo Lazio and Molise National park, central Italy, from 2005 to 2010.

Animal locations ("Pres"=1) were recorded by means of Global Positioning System (GPS) collars, whereas available points ("Pres"=0) were randomly selected at a density of 10 locations/km2 within their seasonal home range to prepresent resources available.

GPS locations were recorded at 3-hr intervals, but their coordinates are omitted from the files due to security reasons (Apennine brown bears are Critically Endangered according to the IUCN Regional Red List Criteria).

The dataset is organized in 8 different Excel files, one for each species and season, as indicated by their names.

The data have been used to conduct K-select seasonal analyses to assess niche overlap between bears and wolves living sympatrically in the central Apennines (Italy).

Data in columns in the Excel comma-separated files:

• "id": Identified or individual bears (F female, M male) or wolf pack (pack-ID)
• "Pres": (1) species presence (used point as from GPS location); (0) random point within individual (bear) or pack (wolf) home range to represent what is available (sampled at 10 locations/km2).

[The following fields pertain to environmental and anthropogenic variables quansitied for both used ("Pres"=1) and available ("Pres"=0) bears' and wolves' locations]

• "Agriculture": % Agricultural areas
• "NoVeg": % Non-vegetated rocky areas
• "PasturesGrassland": %Pastures and grasslands
• "Shrubs": % Shrubland
• "TD_Beech": Average density of trees in beech forest (n° trees/pixel)
• "TD_OakHop": Average density of trees in oak and hop-hornbeam forest (n° trees/pixel)
• "Dist_ForEdge": Distance to forest edges in meters (positive values outside forest, negative inside)
• "Dist_Settl_R1": Distance from urban centers and primary roads (in meters)

• "Dist_R2": Distance from secondary roads (in meters)
• "Hillshade": hillshade
• "TRI": Terrain Ruggedness Index
• "mappa_RSF_SPECIE_STAGIONE": single-specie-RSF (Probability of occurrence of the other species without intraguild predictor)
• "nomevariabile.s": values of the above variables standardized as used in the RSFs.
• "nomevariabile.log": values of the above variables after logarithmic transformation.

Sharing/Access information

Data was derived from the following sources:

• Topographic variables: from a Digital Elevation Model (original resolution 20x20m) provided by the Italian Military Geographic Institute, from which we derived a Terrain Ruggedness Index (TRI), and hillshade (QGIS hillshade, GDAL plugin).
• Land cover data: from regional Corine Land Cover layers (scale 1:10.000) from each administrative region (Lazio, Abruzzo, and Molise). We aggregated the original land cover classes into four categories. From the same land cover data we also calculated the distance to forest edges, with negative values inside the forest and positive outside.
• Tree Cover Density layers: from the Copernicus Land Monitoring Service (https://land.copernicus.eu/en/products/high-resolution-layer-tree-cover-density), we estimated canopy cover and tree density
• Euclidean distance to settlements: calculated from the National Institute of Statistics (2011)
• Euclidean distances to both primary and secondary roads: De Agostini, GeoNext, and TeleAtlas databases (updated to 2003).

All the GIS layers used in the analysis were resampled to a 30x30m cell size. We used a 400m radius moving window around used and available locations and a map-algebra function to transform categorical into continuous variables.

Code/Software

Two sets of R code are included: one referring to the R script used to develp the K-analysis (Ber_wolf_K-select R scripts)ì, and the other to run multi-species Resource Selection Functions (Bear_wolf multispecies RSF_R scripts).