The division of food space among mammalian species on biomes
Data files
Jul 10, 2025 version files 1.11 MB
-
biome_mammals.csv
720.33 KB
-
Food.R
33.76 KB
-
mamm_data_terr.txt
346.48 KB
-
README.md
4.86 KB
Abstract
Understanding how species ecological partition functions across biomes is fundamental to macroecology and conservation biology. Here, we examine the global distribution of dietary strategies in terrestrial mammals, focusing on how biome specialization modulates trophic diversity and structure on a broad geographical scale. Using species-level data from over 3,600 terrestrial mammal species, we constructed a multivariate dietary space and quantified the area, redundancy, dispersion, uniqueness, and turnover of trophic strategies across ten major biomes. Species were classified as biome specialists, moderate generalists, or extreme generalists based on their biome breadth. By analysing biome specialists and generalists separately, we show that biome specialists tend to exhibit more constrained and compositionally distinct dietary niches in less productive biomes, while generalists, particularly moderate generalists, dominate functional space occupancy in all biomes, even the harsher ones such as tundra and taiga. This highlights how environmental constraints and ecological roles shape trophic strategies at global scale. Notably, extreme generalists tended to exhibit more carnivorous or insectivorous diets, suggesting a strategy based on mobile predation or opportunism rather than a highly diversified omnivory. Despite these general patterns, highly productive biomes supported the greatest diversity of dietary strategies, with higher functional redundancy and niche packing. Nestedness and turnover analyses revealed that biome specialists diets are often subsets of generalists diets, but with significant compositional shifts across biomes. These findings underscore the dual role of biome generalists as both functional stabilizers and potential limiters of ecological diversity, and highlight the vulnerability of specialist species to global change. Our study offers a mechanistic framework for understanding how dietary strategies interact with environmental filtering, and for identifying functional risks in changing ecosystems.
Dataset DOI: 10.5061/dryad.mcvdnckc7
Description of the data and file structure
This repository contains an R script used for analyzing the functional trait space of terrestrial mammals in relation to biome specialization and trophic breadth. The analysis includes dimensionality reduction, diversity metrics, redundancy, uniqueness, and statistical testing of functional space occupation across different biome types.
Project Structure
Input Files:
mamm_data_terr.txt
: A tab-delimited text file containing trait and ecological data for terrestrial mammal species. Columns include:
- MSWFamilyLatin: Family name according to Mammal Species of the World (MSW).
- MSW3_ID: Unique species identifier from the MSW3 taxonomy.
- Scientific: Scientific binomial name of the species (e.g., Ornithorhynchus anatinus).
- Diet.Inv: Proportion of diet composed of invertebrates (fuzzy-coded; 0–100 scale).
- Diet.Vend: Proportion of diet composed of vertebrate endotherms (e.g., mammals, birds).
- Diet.Vect: Proportion of diet composed of vertebrate ectotherms (e.g., reptiles, amphibians).
- Diet.Vfish: Proportion of diet composed of fish.
- Diet.Vunk: Proportion of diet composed of unspecified vertebrates.
- Diet.Scav: Proportion of diet composed of scavenged vertebrate material.
- Diet.Fruit: Proportion of diet composed of fruit.
- Diet.Nect: Proportion of diet composed of nectar.
- Diet.Seed: Proportion of diet composed of seeds.
- Diet.PlantO: Proportion of diet composed of other plant matter (e.g., leaves, stems).
- ForStrat.Value: Foraging stratum classification (e.g., “G” = ground-dwelling, “Ar” = arboreal, “S” = scansorial).
- Activity.Nocturnal: Nocturnal activity (binary; 1 = active at night, 0 = not).
- Activity.Crepuscular: Crepuscular activity (binary; 1 = active at dawn/dusk, 0 = not).
- Activity.Diurnal: Diurnal activity (binary; 1 = active during the day, 0 = not).
- BodyMass.Value: Average adult body mass (in grams).
- BodyMass.SpecLevel: Indicates taxonomic resolution of the body mass data (e.g., 1 = species-level, 0 = higher-level estimate).
- litter or clutch size n: Average litter or clutch size (number of offspring per reproductive event).
- order: Taxonomic order of the species (e.g., Monotremata, Didelphimorphia).
biome_mammals.csv
: A CSV file listing mammal species and their biome associations, along with basic taxonomic and ecological traits. Columns include:
- BIOME: Numeric code representing one of 14 terrestrial biomes, following the classification of Hernández Fernández et al. (2022).
- binomial: Scientific name of the species (e.g., Panthera leo).
- presence: Indicates whether the species is present in the given biome (binary; 1 = present, 0 = absent).
- order_: Taxonomic order of the species (e.g., CARNIVORA, RODENTIA).
- family: Taxonomic family (e.g., FELIDAE, MURIDAE).
- genus: Taxonomic genus (e.g., Panthera, Canis).
- terrestial: Indicates whether the species is considered terrestrial (binary; 1 = terrestrial, 0 = aquatic or semi-aquatic).
- Specialist: Biome specialization status based on species distribution breadth (categorical; values include “Specialist” or “Generalist”).
Food.R
: The R script used to perform all analyses described in the related manuscript
Main Analyses Performed
- Principal Coordinates Analysis (PCoA) on fuzzy-coded dietary traits
- Rarefaction Analyses to standardize across sample sizes
- Functional Space Occupation using convex hulls per biome
- Functional Space Overlap Analysis: Nestedness and turnover metrics
- Functional Redundancy across specialization groups and biomes
- Functional Dispersion (FDis) calculation by group
- Functional Uniqueness (nearest-neighbor-based)
- Trophic vs. Geographic Breadth Relationship
- Diet Composition Profiling across biomes and specialization groups
Requirements
This script depends on the following R packages: ade4, dplyr, ggplot2, ggpubr, ggrepel, ggConvexHull, GeoRange, tidyverse, purrr, FD, xlsx, reshape2, scales, nlme, ggeffects, sf.
Usage
- Place your input files in the designated working directory (
setwd()
is specified early in the script). - Run the script in full (
source("Food.R")
) or section by section. - Output files will be generated in
./output2/
and./plots/
.
Notes
- Taxonomic harmonization is performed to align species names across datasets.
- Functional metrics (FDis, uniqueness, redundancy) are computed using the first two PCoA axes.
- Output tables and plots are designed to support publication-ready figures.