Data from: An uncomfortable neighborhood: presence evolution of two competing carnivores in north-eastern Italy
Data files
Nov 13, 2025 version files 108.63 KB
-
GLMM.csv
38.35 KB
-
GLMM.R
3.38 KB
-
MLMM_dataset.csv
62.78 KB
-
MLMM.R
1.13 KB
-
README.md
3 KB
Abstract
Within ecological communities, larger predators typically limit mesocarnivore populations. On a continental scale, this may be the case for the grey wolf (Canis lupus) over the golden jackal (Canis aureus) in Europe. North-eastern Italy represents one of the first areas in Western Europe to experience golden jackal colonization, followed by wolf recolonization. Since few studies have investigated the spatial relationship between these two wild canids, this work aimed to analyze their distribution dynamics and investigate whether the wolf and environmental factors may have influenced golden jackal distribution at a local scale. We used systematic and opportunistic data collected over 11 years (2013-2023) to assess the presence of both species on a 10x10 km square grid system. A multinomial logistic mixed model (MLMM) was applied to test whether the study period and landscape metrics (terrain ruggedness and habitat fragmentation indices) influenced both species’ presence at a broad scale (i.e., square grid units). Generalized linear mixed models (GLMMs) were used to test whether the wolf and landscape metrics influenced jackal presence at a finer scale (i.e., jackal howling calling stations). Both species showed an increase in their former distribution, with a small growth in sympatric areas. Models revealed that jackals preferred less rugged and more fragmented areas typical of the lowland, whereas wolves preferred rugged terrain with extensive forested patches in the Alpine range. Furthermore, our results showed that, at a local scale, golden jackal presence was negatively related to wolf presence. This study provides further insights into the coexistence of these two competing wild canids, suggesting a potential top-down effect of the wolf on golden jackal colonization dynamics. However, the wolf influence may vary in intensity depending on environmental context, with a weaker effect in areas of higher human pressure, such as the lowlands.
Description of the data and file structure
This dataset contains the data and code required to replicate analyses in Frangini et al. (2025)
Files and variables
File: MLMM_dataset.csv
Description: The dataset used to perform the MLMM as described in the manuscript. Each row refers to a single grid square for a specific year.
This .csv file is semicolon-delimited with a comma decimal.
Variables
- ID: the numeric ID of the grid square.
- Species_c1: the presence category for that grid square in a specific year, based on C1 data: "Jackal" or "Wolf", when one of the two species has been found; "Sympatry", when both species have been found; "Absence", when neither jackal nor wolf has been found.
- LDI: the mean value of Landscape Division Index inside the grid square.
- ENN: the mean Euclidean Nearest Neighbor distance among urban patches inside the grid square.
- TRI: the mean value of Terrain Ruggedness Index inside the square grid.
- Area: the biogeographical area where the grid square lies; i.e., within the Alpine or Lowland portion of the study area.
Within the dataset, we mistakenly left the Italian terms: "Alpi" means Alps, and refers to the Alpine part of the study area, while "Pianura" means lowland and refers to the lowland part of the study area. - Year: the sampling year, between 2013 and 2023.
- NP: the number of urban patches inside the grid square.
- Pland: the percentage of urban patches inside the grid square.
File: GLMM.csv
Description: The dataset used to perform the GLMMs as described in the manuscript. Each row refers to a golden jackal howling stimulation in a given year. This .csv file is semicolon-delimited with a comma decimal.
Variables
- Jackal_presence: the golden jackal status at each calling station, present or absent.
- ID_loc: numerical identifier of the jackal howling calling station
- ENN: the mean Euclidean Nearest Neighbor distance among urban patches
- NP: the number of urban patches inside the 2km buffer around the calling station.
- Pland: the percentage of urban patches inside the 2km buffer around the calling station.
- LDI: the mean value of Landscape Division Index inside the 2km buffer around the calling station.
- TRI: the mean Terrain Ruggedness Index value inside the 2km buffer around the calling station.
- Zona: the biogeographical area of the study area, i.e., lowland or alpine
- n_wolves: the number of grey wolf observations collected inside the 2km buffer around each calling station.
- year: the sampling year, i.e., from 2013 to 2023.
Code/software
- "MLMM.R": the R code used to perform the MLMM analysis, as described in the manuscript.
- "GLMM.R": the R code used to perform the GLMMs analyses, as described in the manuscript.
The packages used to perform the analyses are available within the code.