Overview

This repository contains the code and data for the study "Silviculture simplifies anuran–prey networks and increases niche partitioning in the Brazilian Atlantic Forest" by Carvalho et al. (2025).
The study investigates the impact of silviculture practices on the complexity of anuran-prey networks.

Contents

Our dataset has a markdown R script (code_walkthrough.Rmd), raw and processed data (Data/), necessary code and functions (Source/), as well as style (style.css) and bibliography (references.json) files used to generate the knitted version of the R Markdown file (code_walkthrough.html).
All files are contained inside Carvalhoetal_2025_Dryad.zip.

Data Folder

Inside Data/, the following files are included:

Abund.csv

Species abundance data for anurans sampled across multiple sites. Each row represents a sampling site, and each column corresponds to the number of individuals recorded for a given anuran species.

All_trail_info.csv

Information about sampling sites.

• Trail : Sampling site code
• Environment : Environment where the specimen was sampled
  • eucalipto : Eucalyptus plantation
  • mata : Atlantic Forest fragment
• Lat : Latitude in decimal degrees (WGS84)
• Long : Longitude in decimal degrees (WGS84)
• Period : Sampling period for each transect
• Bouts : Number of bouts for anuran sampling

Connectance_Lit_Data.csv

File containing data on previously published anuran–prey networks.

• Id : Identification for original references, as cited in Carvalho et al. (2025) Supplementary Information
• Country : Country where the original study site is located
• Lat : Latitude in decimal degrees (WGS84)
• Long : Longitude in decimal degrees (WGS84)
• Ranuran : Number of anuran species (anuran nodes) in the network
• Rprey : Number of prey (prey nodes) in the network
• Size : Number of nodes (Ranuran + Rprey) in the network
• Conn : Connectance values for each network. Values range from 1 (fully connected network) to 0 (fully disconnected network)

Data_filt_MW_SVL.csv

Raw dietary data.

• Species : Scientific names of each anuran species from which each prey item recorded was identified
• Code : Unique field number of each specimen from which each prey item recorded was identified
• MW : Mouth width (mm) of the specimen from which each prey item recorded was identified
• SVL : Snout-vent length (mm) of the specimen from which each prey item recorded was identified
• Weight : Weight (g) of the specimen from which each prey item recorded was identified
• Trail : Sampling site code
• Environment : Environment where the specimen was sampled
  • eucalipto : Eucalyptus plantation
  • mata : Atlantic Forest fragment
• Order : Taxonomic classification of prey item at the Order level. If Empty, stomach contained no identifiable prey item
• Class.inferior : Taxonomic classification of prey items to the lowest level possible, usually suborder. If Empty, stomach contained no identifiable prey item
• Quantity : Number of prey items with the same taxonomic classification and morphotype encountered for the same specimen
• Length : Length (mm) of the recorded prey item
• Width : Width (mm) of the recorded prey item
• Obs : Column used for plain text observations
• Volume : Total prey item volume as derived from the ellipsoid formula and multiplied by Quantity
• Ind.Volume : Individual prey item volume as derived from the ellipsoid formula (4 * pi * Length * Width^2)/3

Diet_sample_indv_vol.csv

Dietary data summarized using volume and specimen identification.

• Species : Scientific names of each anuran species from which each prey item recorded was identified
• Code : Unique field number of each specimen from which each prey item recorded was identified
• Environment : Environment where the specimen was sampled
  • eucalipto : Eucalyptus plantation
  • mata : Atlantic Forest fragment
• Remaining columns represent prey item categories and the summed volume for each specimen–prey type combination

Distance_matrix.csv

Distance matrix (km) between sampling sites.

Freq_eu_diet_matrix_final.csv

Weighted interaction matrix describing trophic relationships between anuran predators (columns) and their prey (rows) in the Eucalyptus environment. Cell values correspond to the frequency of occurrence of each prey type in the diet, representing the strength of each predator–prey interaction.

Freq_mt_diet_matrix_final.csv

Weighted interaction matrix describing trophic relationships between anuran predators (columns) and their prey (rows) in the Atlantic Forest environment. Cell values correspond to the frequency of occurrence of each prey type in the diet, representing the strength of each predator–prey interaction.

PA_MV.csv

Raw prey availability data.

• Trail : Sampling site code
• Environment : Environment where the specimen was sampled
  • eucalipto : Eucalyptus plantation
  • mata : Atlantic Forest fragment
• Order : Taxonomic classification of prey item at the Order level. If Empty, stomach contained no identifiable prey item
• Class.inferior : Taxonomic classification of prey items to the lowest level possible, usually suborder. If Empty, stomach contained no identifiable prey item
• Length : Length (mm) of the recorded prey item
• Width : Width (mm) of the recorded prey item
• Volume : Individual prey item volume as derived from the ellipsoid formula (4 * pi * Length * Width^2)/3

PA_sample_composition_freq.csv

Prey availability data summarized using frequency.

• Environment : Environment where the specimen was sampled
  • eucalipto : Eucalyptus plantation
  • mata : Atlantic Forest fragment
• Trail : Sampling site code
• Remaining columns represent prey item categories and the summed frequency for each site–prey type combination

PA_sample_composition_vol.csv

Prey availability data summarized using volume.

• Environment : Environment where the specimen was sampled
  • eucalipto : Eucalyptus plantation
  • mata : Atlantic Forest fragment
• Trail : Sampling site code
• Remaining columns represent prey item categories and the summed volume for each site–prey type combination

Rev1_items_sens.csv

A summary of all diet specialization metrics measured for anurans across environments.

• Species : Scientific names of each anuran species from which specialization metrics were derived
• Environment : Environment where the specimen was sampled
  • EU : Eucalyptus plantation
  • AF : Atlantic Forest fragment
• IT : Number of prey items used for analyses
• Z : Z-score for the dissimilarity between 1000 random diet and the observed environmental availability, compared to the dissimilarity between the observed diet and environmental availability. An anuran is considered to be foraging selectively if the observed dissimilarity was significantly greater than that from the simulated diets (i.e., Z-score > 2)
• sp50 : Minimum number of prey required to account for fifty percent of the diet.
• D : Electivity for the primary prey.

Vol_eu_diet_matrix_final.csv

Weighted interaction matrix describing trophic relationships between anuran predators (columns) and their prey (rows) in the Eucalyptus environment. Cell values correspond to the volume of each prey type in the diet, representing the strength of each predator–prey interaction.

Vol_mt_diet_matrix_final.csv

Weighted interaction matrix describing trophic relationships between anuran predators (columns) and their prey (rows) in the Atlantic Forest environment. Cell values correspond to the volume of each prey type in the diet, representing the strength of each predator–prey interaction.

Source

Inside Source/, there are five .R scripts containing necessary code or functions used by the file code_walkthrough.Rmd. The following files are included:

bootstrap_diet_vol.R

Includes a function that executes a bootstrap to simulate null model networks. From the total pool of dietary items (prey items consumed), we randomly selected prey items without replacement and assigned them to each species, respecting the observed number of prey items per species and total volume per prey type.

diet_gen.R

Function to generate diets based on a given environment (prey abundance).

morans.R

Function to organize the calculation of Moran's I statistic and hypothesis testing.

quant_to_count.R

Function to expand a data frame based on a quantity column (here prey items). This function duplicates rows in the input data frame based on the value in the specified column, so that each unit of the quantity is represented as a separate row.

sp50.R

Function to calculate the SP50 for a diet abundance distribution. SP50 is the point where the cumulative proportion of the diet reaches 0.5 (or another specified level).

Requirements

• R (version 4.0 or higher)
• R packages: dplyr, ggplot2, tidyverse, vegan, igraph, bipartite, ggplot2, parallel, iNEXT

Acknowledgements

The authors thank all those who helped to collect the data during field campaigns. We thank Karoline Ceron for sharing unpublished data. Also, we thank Melhoramentos Ltda. for supporting this project.
ANC was funded by São Paulo Research Foundation – FAPESP (grant: #2023/06286-7 and 2023/16700-3).
MM thanks Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for a research fellowship (#309772/2021-4).
PRG was supported by FAPESP (grant: #2018/14809-0) and CNPq.
Fieldwork was possible with funding from FAPESP (grants: #2018/14091-1 and #2020/12658-4).