Data from: Forest edges buffer effects of agricultural management on bee–plant interaction networks
Data files
Jul 06, 2026 version files 59.54 KB
Abstract
Bee–plant interaction networks are essential for sustaining pollination services in agricultural landscapes, and farming practices can modify their structure by altering species composition and interaction patterns along crop margins.
This study assessed how conventional and organic management influence bee–plant interaction networks and key network metrics (connectance, nestedness, interaction richness, interaction diversity, and modularity) along crop field edges bordering native forest fragments.
Surveys conducted on farms growing different crops revealed similar bee and plant species richness between management systems. However, conventionally managed farms were characterized by a greater dominance of supergeneralist species, such as Apis mellifera Linnaeus, 1758, and Trigona spinipes (Fabricius, 1793), resulting in higher network connectance. In contrast, organically managed farms exhibited greater interaction richness and interaction diversity but lower connectance, indicating a broader distribution of interactions among species rather than a greater number of interacting taxa. The presence of flowering plants along crop borders in both systems promoted active bee–plant networks regardless of management type.
Our results highlight the importance of crop edges and adjacent native vegetation in maintaining pollination processes across agricultural systems. They also suggest that a shared landscape context and similar floral resources can reduce expected differences between organic and conventional management, thereby mitigating potential negative effects of conventional practices on pollination networks.
Dataset DOI: 10.5061/dryad.p5hqbzm3n
Description of the data and file structure
This dataset contains bee–plant interaction records collected along crop field edges adjacent to native forest fragments in agricultural landscapes of southeastern Brazil. The data include sampling site information, agricultural management type (organic or conventional), bee and plant taxonomic identifications, bee functional traits, geographic coordinates, and sampling dates. These data were used to construct bee–plant interaction networks and to calculate network metrics, including connectance, nestedness, interaction richness, interaction diversity, and modularity, to evaluate the effects of agricultural management on pollination networks.
Files and variables
File: Forest_Edges_Buffer_Agricultural_Management_Effects_on_Bee–Plant_Interaction-dataset2.xlsx
Description:
Variables
-
CODE: unique identifier for sampling unit or record -
Management: agricultural management system (conventional or organic) -
Latitude: latitude of sampling location (decimal degrees) -
Longitude: longitude of sampling location (decimal degrees) -
Location: sampling site identifier. The abbreviations correspond to the following sampling sites.Conventional management
- COEO: Sítio Ogata, Biritiba-Mirim (Rio Acima).
- COSK: Sítio Kajihara, Mogi das Cruzes (Ponte Grande).
- COCH: Cooperativa Hasegawa, Biritiba-Mirim (Rio Acima).
- COAM: Sítio Adm Mauro, Salesópolis (Nossa Senhora dos Remédios).
- COSTD: Sítio Tadeu e Regina, Mogi das Cruzes (Conjunto Santo Ângelo).
- COFR: Sítio Renan, Biritiba-Mirim (Rio Acima).
Organic management
- OMO: Sítio Orgânico Moatan, Mogi das Cruzes (Taiaçupeba).
- OOS: Sítio Orgânico da Serra, Mogi das Cruzes (Biritiba-Ussu).
- OV: Sítio Victor, Suzano (Ipêlândia).
- OAK: Sítio Akinaga, Biritiba-Mirim (Centro).
- OJO: Sítio José, Salesópolis (Mirante).
- OSI: Sítio Imada, Mogi das Cruzes (Cocuera).
-
Family: bee family -
Subfamily: bee subfamily -
Tribe: bee tribe -
Genus: bee genus -
Species: bee species (scientific name) -
Sociality: level of social organization (e.g. solitary, social, eusocial) -
Size_classification: categorical body size class -
size: continuous body size measurement (unit: mm, if applicable) -
Nesting: bee nesting habit (e.g. above ground or ground) -
Feeding specialization: degree of dietary specialization of bee species -
Plant_family: plant family visited -
Plant_species: plant species visited (scientific name) -
Sampling_date: date of sampling (YYYY-MM-DD) -
year: year of sampling
Code/software
All analyses were performed in R (version 4.3.0 or later) using open-source packages. No proprietary software was used.
R packages used
- tidyverse
- readr
- dplyr
- ggplot2
- vegan
- bipartite
- igraph
Workflow
- Import of raw CSV data using readr
- Data cleaning and standardization of taxonomic and trait information
- Construction of bee–plant interaction matrices
- Calculation of network metrics (connectance, nestedness, interaction richness, interaction diversity, modularity)
- Statistical analyses comparing agricultural management systems
- Visualization of results
All datasets are linked through shared identifiers such as CODE, Species, and Plant_species.
Access information
Other publicly accessible locations of the data:
No additional publicly accessible repositories or databases contain these data. All data are available only through this Dryad submission.
Data was derived from the following sources:
No external datasets were used. All data were generated from field sampling conducted in agricultural landscapes in southeastern Brazil.
