Living under the scope: behavior affects survival in a heavily harvested and long-lived ungulate
Data files
Mar 13, 2025 version files 427.61 MB
-
all_plots_data.rds
79.44 MB
-
high.rds
82.53 MB
-
hsa.R
28.42 KB
-
low.rds
82.46 MB
-
pre-rut.rds
92.61 MB
-
README.md
3.92 KB
-
rut.rds
90.54 MB
-
survival_mod.R
4.11 KB
Abstract
The spatiotemporal behavior of game species may play a critical role in their survival throughout the hunting season. Where humans are the most dominant predators, avoidance of landscape features that allow hunters access to hunting grounds can be key to increasing survival. However, in Sweden there is limited knowledge about how hunters impact populations through selection of behavioral traits. In this system, hunting pressure is generally high, and approximately 30% of the moose (Alces alces) population is harvested per year. To test for differences in habitat selection in relation to survival under varying levels of hunting pressure in this long-lived game species, we analyzed 10 years of individual-based movement data of adult male and female moose in southern Sweden. We used integrated Step Selection Functions to analyze habitat selection in four consecutive time periods (before the rutting season, during the rut, during the first three weeks of the hunting season and the following three weeks). We matched habitat selection coefficients of individual animals with their fate and tested for behavioral differences between harvested and non-harvested male and female moose in each period. Our findings indicate that hunters may select for sex-specific behavioral traits in habitat selection by adult moose. We found that hunters harvested male moose that selected for higher NDVI during the rut, suggesting personality-driven hunting mortality in male moose. In contrast, in female moose, the mortality risk increased with selection for shorter distance to roads during the hunting season, suggesting selection for behavioral traits by hunters. Our findings indicate that current harvest practices for moose result in a selection for sex-specific behavioral traits. Our study highlights the complex dynamics of survival in a long-lived species under high hunting pressure, revealing how sex-specific habitat selection may impact mortality risk by hunters and, in return, might influence population management.
Dataset DOI: 10.5061/dryad.ns1rn8q47
This dataset includes 2 R-scripts and 5 datasets to perform the analyses conducted in the accompanying paper. All scripts are run on R version 4.4.1
All scripts automatically call the needed respective datasets.
Description of the data and file structure
File: hsa.R
Description: Script to run habitat selection analysis. Call the datasets: all_plots_data.rds, high.rds, low.rds, pre-rut.rds and rut.rds
File: all_plots_data.rds
Description: fitted iSSF for all animals across all years.
Used in scripts
- OBJECT_ID = animal collar ID
- Sex = sex of the animal (male/female)
- Hunt_year = the year in which the hunting season could take place. Equal to year, used to match NDVI and dist_forest
- ta_ = turning angles between GPS positions
- costa_ = cosine transformed ta_
- sl_ = meters between GPSpositions
- log_sl_= log transformed sl_
- case_ = TRUE/FALSE, observed or random step
- Age = estimated (by toothwear) age of the animal in years
- y = numeric of case_ (1,0)
- unique_step_id = unique step identifier, combination of case and step_id_ (step_id_ has been removed from the dataset)
- road_dist = distance to roads in m at every GPS position
- z_road_dist = z-score transformed road_dist at every GPS position
- dist_forest = distance to forest in m at every GPS position
- z_dist_forest = z-score transformed dist_forest at every GPS position
- NDVI = Normalized Difference Vegetation Index at every GPS position
- fate_ultimal = ultimate fate of animal (harvest, died of other causes, or still alive)
- remaining_lifetime = remaining lifetime until death date in years
- relative_days_to_start = used to define the days relative to the start of the hunting season, necessary as start of hunting season shifts every year
- hunting_intensity = factor w/ 3,classes used to subset data into: high pressure, low pressure and no pressure. later reclassifies "no pressure" into pre-rut and rut in the variable hunt_intes
- hunt_intes = factor w/ 4,classes used to subset data into: high pressure, low pressure, pre-rut and rut (based on hunting_intensity). Used to split the data into subsets for modelling
not used in scripts
- ID = unique column identifier, not used in scripts
- tod_end_ = time of day w/ two levels (day or night)
- day_night = tod_end, but different calculation of time of day
- fate_season = fate in specific hunting season part
- fate_huntyear = fate in respective hunting season
- (z_)forest_cohesion_[250,500,1000]m: Landscape metric, not used in script. Z-score transformed
- (z_)slope = slope at GPS - position in degree
- (z_)-tri= terrain ruggedness index
- hour / month = hour or month of specific GPS position, not used
- logmhr = log-transformed sl_ divided by 3h (sampling rate of GPS data), distance travelled per hour in m
- landuse_end = landuse class according to the national landcover map of sweden
- name = data definition of landuse_end
File: high.rds
Description: best model for high hunting pressure after AIC - model selection. Contains a standard glmmTMB - model.
File: low.rds
Description: best model for low hunting pressure after AIC - model selection. Contains a standard glmmTMB - model.
File: pre-rut.rds
Description: best model for the pre- rut
File: rut.rds
Description: best model for the rut after AIC - model selection. Contains a standard glmmTMB - model.
File: survival_mod.R
Description: performs the analysis on survival in moose. Run on R-version 4.4.1
Code/software
The necessary scripts are included in the data and are run on R version 4.4.1## Access information
Other publicly accessible locations of the data:
- NA
Data was derived from the following sources:
- NA
