Data from: Reaching reproduction in a large carnivore: Role of early environment and intrinsic traits
Data files
Sep 26, 2025 version files 81.86 KB
-
Analysis_script.R
17.57 KB
-
reaching_reproduction_frame.csv
60.97 KB
-
README.md
3.32 KB
Abstract
To reach reproduction, individuals must survive the juvenile stage, a critical period of low survival rates in large carnivores. Early-life conditions during this stage can have lasting effects on survival, reproductive maturation, growth, physiology, and behaviour. We assessed recruitment probability in Scandinavian wolves, i.e., the probability that a wolf to reaches the reproductive stage and has pups surviving at least five months of age. To unravel human-related and biological factors within the natal territory that could affect recruitment probability, we analysed life-history data from 582 Scandinavian wolves (Canis lupus) identified by DNA as pups or juveniles in their birth territory. Factors considered included main prey density, road density, human density, and proximity to non-breeding zones, as well as sex, inbreeding level and collaring. Among the 582 wolves analysed, 122 produced at least one surviving pup, corresponding to a recruitment probability of 0.21. Recruitment probability was more than twice as high (0.5) for juvenile wolves fitted with GPS collars compared to non-collared individuals (0.22), and was positively correlated with human population density in the natal territory. We found no significant effects of other biologically or human-related predictors. These results suggest that in this large carnivore population, managed below carrying capacity, individual recruitment probability is primarily influenced by human-related factors, potentially reflecting poaching risk.
Dataset: reaching_reproduction_frame.csv
File Format:
Name: reaching_reproduction_frame.csv
Type: CSV (Comma-Separated Values)
Delimiter: ,
Description:
The dataset reaching_reproduction_frame.csv contains the folowing information about individual wolves:
-Reproduction (categorical): Whether the individual reached reproduction (1 = Yes, 0 = No).
-Collared (categorical): Whether the individual was collared for tracking (Yes, No).
-Inbreeding (numeric): Inbreeding coefficient of the individual based on the pedigree.
-Sex (categorical): Sex of the individual (M = Male, F = Female).
-Human_density (numeric): Density of humans in the individual's natal territory in number of inhabitants per km2.
-Average_gravel_road (numeric): Average gravel road density in the natal territory in km per km2.
-Country (categorical): Birth country of the individual: location of the centroid of the natal territory (S = Sweden, N = Norway)
-Wolf_density (numeric): Number of bordering neighbouring territories, i.e. the number of territories overlapping with the natal territory.
-Dist_non_breeding_zone (numeric): Distance in m between the centroid of the natal territory and the nearest non-breeding zone, i.e. area where the wolves were not allowed to establish.
-Snow_depth (numeric): Average snow depth in cm in the natal territory during cold month: from the 1st of May (average birth of pups) to 30th of April next (standardized).
-Moose_density (numeric): Hunting bag records of moose counted as killed/10 km2 for counties. Data on hunting bag records was generated as a weighted average of the moose density of the counties overlapping with the natal territory.
-Legally_killed (binary): Whether the individual was legally killed (1 = Yes, 0 = No).
-Date of death (numeric): Date of death for the individual, if known; otherwise recorded as "null". Dates are stored in numeric format, with origin set to "1899-12-30".
-Birth_year (numeric): Year of birth of the individual.
-Parental_pair_id (categorical): Identifier for the parental pair.
-Age_known (binary): Whether the age of the individual is known with certainty or estimated (1 = Yes, 0 = No).
Script: Analysis of Factors Influencing Reproduction
File Format:
Name: Analysis_script.R
Type: R
Purpose:
The R script provided analyzes performed in the paper "Reaching reproduction in a large carnivore: role of early environment and intrinsic traits". It includes:
Data Preparation:
Standardizing numerical predictors.
Converting variables to appropriate data types (e.g., factors for categorical variables).
Filtering data to create subsets for specific analyses (e.g., excluding legally killed or collared individuals).
Model Building:
Generalized linear mixed models (GLMMs) using the glmmTMB package.
Fixed effects include environmental, genetic, and anthropogenic variables.
Random effect: Parental_pair_id.
Diagnostics and Model Checking:
Checking multicollinearity using check_collinearity (from the performance package).
Residual diagnostics with the DHARMa package.
Model Selection and Averaging:
Using MuMIn::dredge to identify top models based on AICc.
Averaging models within a delta AICc of 2.
Visualization:
Correlation matrices visualized with the corrplot package.