Data from: Numerical top-down effects on red deer (Cervus elaphus) are mainly shaped by humans rather than large carnivores across Europe
Data files
Oct 05, 2023 version files 1.94 MB
Abstract
Terrestrial ecosystems are shaped by interacting top-down and bottom-up processes, with the magnitude of top-down control by large carnivores largely depending on environmental productivity. While carnivore-induced numerical effects on ungulate prey populations have been demonstrated in large, relatively undisturbed ecosystems, whether large carnivores can play a similar role in more human-dominated systems is a clear knowledge gap. As humans influence both predator and prey in a variety of ways, the ecological impacts of large carnivores can be largely modified. We quantified the interactive effects of human activities and large carnivore presence on red deer (Cervus elaphus) population density and how their impacts interacted and varied with environmental productivity
Data on red deer density were collected based on a literature survey encompassing 492 study sites across 28 European countries. Variation in density across study sites was analysed using a generalised additive model in which productivity, carnivore presence (grey wolf, European lynx, Brown bear), human activities (hunting, intensity of human land-use activity), site protection status and climatic variables served as predictors.
The results showed that a reduction in deer density only occurred when wolf, lynx and bear co-occurred within the same site. In the absence of large carnivores, red deer density varied along a productivity gradient without a clear pattern. Although a linear relationship with productivity in the presence of all three large carnivore species was found, this was not statistically significant. Moreover, hunting by humans had a stronger effect than the presence of all large carnivores in reducing red deer density and red deer density increased with increasing intensity of human land-use, with stronger large carnivore effects (all three carnivore species present) at sites with low human land-use activities.
Synthesis and applications: This study provides evidence for the dominant role played by humans (i.e. hunting, land-use activities) relative to large carnivores in reducing red deer density across European human-dominated landscapes. These findings suggest that when we would like large carnivores to exert numeric effects, we should focus on minimizing human impacts to allow the ecological impacts of large carnivores on ecosystem functioning.
README: Numerical top-down effects on red deer (Cervus elaphus) are mainly shaped by humans rather than large carnivores across Europe
https://doi.org/10.5061/dryad.0cfxpnw7w
Suzanne T.S van Beeck Calkoen, Dries P.J. Kuijper, M. Apollonio, Lena Blondel, Carsten F. Dormann, Ilse Storch, Marco Heurich
Abstract
- Terrestrial ecosystems are shaped by interacting top-down and bottom-up processes, with the magnitude of top-down control by large carnivores largely depending on environmental productivity. While carnivore-induced numerical effects on ungulate prey populations have been demonstrated in large, relatively undisturbed ecosystems, whether large carnivores can play a similar role in more human-dominated systems is a clear knowledge gap. As humans influence both predator and prey in a variety of ways, the ecological impacts of large carnivores can be largely modified. We quantified the interactive effects of human activities and large carnivore presence on red deer (Cervus elaphus) population density and how their impacts interacted and varied with environmental productivity
- Data on red deer density were collected based on a literature survey encompassing 492 study sites across 28 European countries. Variation in density across study sites was analysed using a generalised additive model in which productivity, carnivore presence (grey wolf, European lynx, Brown bear), human activities (hunting, intensity of human land-use activity), site protection status and climatic variables served as predictors.
- The results showed that a reduction in deer density only occurred when wolf, lynx and bear co-occurred within the same site. In the absence of large carnivores, red deer density varied along a productivity gradient without a clear pattern. Although a linear relationship with productivity in the presence of all three large carnivore species was found, this was not statistically significant. Moreover, hunting by humans had a stronger effect than the presence of all large carnivores in reducing red deer density and red deer density increased with increasing intensity of human land-use, with stronger large carnivore effects (all three carnivore species present) at sites with low human land-use activities.
- Synthesis and applications. This study provides evidence for the dominant role played by humans (i.e. hunting, land-use activities) relative to large carnivores in reducing red deer density across European human-dominated landscapes. These findings suggest that when we would like large carnivores to exert numeric effects, we should focus on minimizing human impacts to allow the ecological impacts of large carnivores on ecosystem functioning.
Keywords: Cervus elaphus, top-down control, numerical effects ,large carnivores, environmental productivity, hunting by humans, human land-use activities
Extract files
To extract all Data and Rscripts, download the zip file and unpack on your own device.
Rproject - Running the Rscripts
To run all scripts, open the R-project "SvBC_2023_RedDeer".
Besides the R-project, you will see the folders "Data", "Rscripts" and "Results" and an additional Rmarkdown script called "Concurvity".
To run all analyses described within the manuscript, open the R project and run each of the Rscripts from the folder "Rscripts". All data necessary to run these scripts are stored in the "Data" folder and all R outputs and graphs will be saved in the folder "Results".
The Rmarkdown script "Concurvity" contains the results of the concurvity checks conducted as part of the model diagnostics. A description of this is included within the Supplementary Information and the Rcode is included within Rsript "04. SuppInfo_III".
Data
1. Data_SvBC_Reddeer
Dataset containing all variables extracted from the literature search and additional factors shown to affect ungulate density as explained within the "Material and Methods: Variables potentially influencing red deer density" within the manuscript and Supporting information.
- Variables literature search: Year study published\, country\, study area name\, latitude\, longitude\, red deer density\, hunting.
- Additional factors shown to affect ungulate density: net primary productivity\, large carnivore presence\, human influence index\, protection status of the study site\, percentage of tree cover\, Palmer drought severity index\, Normalised difference snow index.
2. Data_SvBC_Reddeer_method
Additional dataset with a specification of which methods were used for red deer density estimation as described in the studies included in the literature search. The large variation in methods is discussed within the "Discussion: Large carnivore impacts on red deer density across Europe".
3. Supp_SvBC_RedDeer_5km
Dataset containing all variables as discussed under "1. Data_SvBC_Reddeer" where each of the variables are calculated based on a 5-km buffer instead of the 10-km buffer as included within the manuscript. The results of our sensitivity analyses, including the model with the variables based upon a 5 km-buffer, are summarized in the Supporting Information. The Rscript for these analyses can be found in the Rscript "SuppInfo II".
4. Folder spatial
Includes the datasets for each variable calculated with a 5 km-buffer as explained above. These datasets are used to run the Rscript "SuppInfo II".
Rscript structure
01. Model
- Data preparation
- Pre-modelling exploratory analyses
- Model
- Post-modelling diagnostics
- Results
02. Figures
- Repetition model
- Boxplots parametric coefficients
- Plot smooths
- Human influence index * Predation
- Net primary productivity * Predation
- Environmental plots (tree canopy cover, Normalized difference snow index , Palmer drough severity index)
- Figures
03. SuppInfo_II
- 5-km buffer
- Data preparation
- Pre-modelling exploratory analyses
- Model
- Post-modelling diagnostics
- Results
- 5000 x 5000 resolution
- Data preparation
- Pre-modelling exploratory analyses
- Model
- Post-modelling diagnostics
04. SuppInfo_III
- Check stability model, randomly deleting 10% of the data
- Data preparation
- Model
- Post-modelling diagnostics
- Results
- Figures
- Randomly change density estimates by 60%
- Data preparation
- Model
- Post-modelling diagnostics
- Results
- Figures
- Delta deviance explained
Methods
Data on red deer density were collected based on a literature survey encompassing 492 study sites across 28 European countries. Variation in density across study sites was analysed using a generalised additive model in which productivity, carnivore presence (grey wolf, European lynx, Brown bear), human activities (hunting, intensity of human land-use activity), site protection status and climatic variables served as predictors.