Overabundant populations of large wild herbivores disrupt plant-pollinator networks in a Mediterranean ecosystem
Data files
Jun 20, 2025 version files 26.46 KB
-
database_Hernández-Castellano_et_al_2025_Plant_Biology.zip
23.51 KB
-
README.md
2.95 KB
Abstract
Large herbivores are keystone species, so changes in their population abundance may cause cascading effects on ecosystems. Our study explores the impacts of increasing red deer (Cervus elaphus) densities on plant-pollinator networks. We conducted a manipulative experiment with increasing densities, consisting of comparable hectare-scale enclosures in a Mediterranean ecosystem. We simulated two current scenarios of deer overabundance: high densities (>30 individuals/km2) and hyper densities (>90 individuals/km2). We compared these scenarios to an adjacent control exclosure (no deer).
In each scenario we analysed herbivory in three focal species of shrub, as well as flower and pollinator abundance, richness, diversity, evenness, and composition. We also analysed plant-pollinator interaction dissimilarity among networks, plant visitation rate, as well as plant and pollinator generality. For each network, we analysed network size, interaction richness, diversity, and evenness, as well as network connectance, specialization, nestedness, modularity, modules, and robustness.
We found that herbivory reduced flower abundance of scrubs, as well as flowering plant and pollinator richness. Remaining plants and pollinators lost interactions, and some plants lost pollinator visitors. Network specialization and modularity decreased because modules (groups of species strongly connected) conformed by herbs and specialist pollinators were gradually extirpated as deer density increased. This simplification increased network connectance and nestedness. Network robustness (a measure of stability) remained unaltered because the dominant plant, which attracted most pollinators, was unpalatable to deer.
We conclude that in overabundant deer scenarios: 1) impact on plant-pollinator networks will increase with deer density; 2) plant-pollinator networks will be eroded, especially if they are composed of palatable, rare plants, visited by specialist pollinators; and 3) plant-pollinator network stability will be unaffected if dominant plants are generalists and unpalatable to deer.
All this information enables to replicate all the analyses present in the manuscript, as well as to describe or explore other aspects of the study system.
Please, have a look in the README file before exploring or using the data.
If you have any questions, please contact the author.
This dataset is part of a long-term (>4 years) experiment, so may be updated in the forthcoming years. If you are interested in using the data of this study or the data of the long-term experiment for your research, please contact the author. I will be happy to collaborate.
Dataset DOI: 10.5061/dryad.tqjq2bw67
Description of the data and file structure
We conducted a manipulative experiment with increasing densities, consisting of comparable hectare-scale enclosures in a Mediterranean ecosystem. We simulated two current scenarios of deer overabundance: high densities (>30 individuals/km2) and hyper densities (>90 individuals/km2). We compared these scenarios to an adjacent control exclosure (no deer).
In each scenario we analysed herbivory in three focal species of shrub, as well as flower and pollinator abundance, richness, diversity, evenness, and composition. We also analysed plant-pollinator interaction dissimilarity among networks, plant visitation rate, as well as plant and pollinator generality. For each network, we analysed network size, interaction richness, diversity, and evenness, as well as network connectance, specialization, nestedness, modularity, modules, and robustness.
Files and variables
All the databases of the study are present here:
-“herbivory.txt”. herbivory on the three focal plants (semiquantitative measure (range 0-5) following Perea et al. 2015)
-“plant_community.txt”. plant community composition (flower abundance of each species)
-“plant individuals.txt” number of individuals of the two subdominant focal plants
-“pollinator_community.txt”. pollinator community composition (number of individuals of each species)
-“visitation rate.txt”. visitation rate of the three focal plants (observed pollinator individuals/100 observed flowers)
-plant-pollinator networks, 3 for each treatment (“nw_qc1.txt”, “nw_qc2.txt”, “nw_qc3.txt”, “nw_qdb1.txt”, “nw_qdb2.txt”, “nw_qdb3.txt”, “nw_qda1.txt”, “nw_qda2.txt”, “nw_qda3.txt”). Each network is an adjacency matrix in which files are plants, columns are pollinators, and cell value is interaction frequency (standardised number of visits a given pollinator visited a given plant).
-the metanetwork for each treatment (3 local networks pooled together; “metaweb_C.txt”, “metaweb_dB.txt”, “metaweb_dA.txt”)
-the metanetwork for the whole study system (3 metanetworks pooled together; “metaweb.txt”)
All this information enables to replicate all the analyses present in the manuscript, as well as to describe or\
explore other aspects of the study system.
Please note that some codes and abbreviations are used in the databases for convenience:
TREATMENT CODES
C: Control\
dB: High Density\
dA: Hyper Density
SPECIES ABBREVIATIONS
Clad: Cistus ladanifer\
Cpop: Cistus populifolius\
Tmas: Thymus mastichina
Usage note: You can easily open this files with Excel or R (R Core Team 2025) to conduct descriptive statistics. You may use the package vegan (Oksanen et al. 2025) to visualize community data, and the package bipartite (Dormann 2025) to visualize networks.