Data and code from: Incorporating pollinator movement into connectivity models predicts pollen-mediated gene flow and highlights the importance of regenerating forests in tropical landscapes
Data files
Mar 24, 2026 version files 35 MB
-
DRYAD_Clean.zip
34.99 MB
-
README.md
14.37 KB
Abstract
Pollen-mediated gene flow is crucial for ecological and evolutionary processes, and understanding its disruption by anthropogenic disturbances is essential for conservation. In this study, we developed landscape connectivity metrics that incorporated hummingbird movement behaviour to assess how structural (amount and configuration) and functional (species-specific behavioural response) landscape connectivity influences pollen-mediated gene flow in a tropical plant species. We adapted the incidence function model (IFM) to develop a set of functional landscape connectivity metrics that integrated pollinator movement behaviour. We evaluated whether these metrics outperform structural landscape connectivity metrics for explaining contemporary pollen-mediated gene flow. The performance of landscape connectivity metrics as predictors of contemporary pollen-mediated gene flow is scale dependent. At the focal patch scale, pollen-mediated gene flow was better predicted by structural connectivity metrics, specifically the area of contiguous mature forest. At the local landscape scale, pollen-mediated gene flow was better predicted by functional connectivity metrics that accounted for hummingbird movement, including gap-crossing probabilities. We found that including regenerating forests and narrow forest elements better explained pollen-mediated gene flow than a focus solely on mature forest. By integrating hummingbird movement behaviour, we offer a more realistic and nuanced understanding of how landscapes influence pollen-mediated gene flow. We underscore the importance of accounting for pollinator movement behaviour in conservation strategies that aim to preserve pollen-mediated gene flow, and demonstrate that forest regeneration in critical for maintaining functional landscape connectivity in tropical fragmented landscapes.
Dataset DOI: 10.5061/dryad.3ffbg7b08
Description of the data and file structure
This dataset was collected to evaluate how structural and functional components of landscape connectivity influence contemporary pollen-mediated gene flow in Heliconia tortuosa (Heliconiaceae), a hummingbird-pollinated tropical plant. We developed landscape connectivity metrics by adapting the Incidence Function Model to integrate habitat configuration with empirical data on pollinator movement behaviour, including daily flight distance, home range length, and gap-crossing probability. Connectivity was quantified at two spatial scales: within focal habitat patches and across the surrounding 1 km landscape (corresponding to the maximum daily movement range of hummingbirds). We evaluated alternative definitions of trapliner habitat to test the relative importance of mature forest, regenerating forest, and narrow forest elements for maintaining functional connectivity. Contemporary pollen-mediated gene flow was quantified using previously published genetic data, and alternative structural and functional connectivity metrics were compared to assess their explanatory power. The experimental effort provides a landscape genetic framework for identifying the components of connectivity most critical for sustaining pollen transport and long-term population viability in fragmented tropical landscapes.
Repository Structure
The repository is organized into three top-level folders:
- Code/
- Data/
- Output/
1. Code/
This folder contains all R scripts used to integrate hummingbird movement parameters into functional connectivity metrics:
- Estimate structural and functional landscape connectivity metrics (adapted from the Incidence Function Model).
Level_Framework_Analysis.Rmdcontains the main landscape connectivity analyses conducted in the study. - The
Patch_Based_Metrics.Rcontains code to estimate landscape connectivity metrics restricted to the focal forest patches. - The
Whole_Landscape_Metrics.Rcontains code to estimate landscape connectivity metrics not restricted to the focal forest patches. - Generate a series of simulated landscapes (
SI_Simulations). This includes scripts that create a series of simulated landscapes used to evaluate the behaviour and scaling properties of the connectivity metrics. This subfolder contains raster files generated during simulation analyses. - Calculate connectivity metrics for:
- The focal patch only
- The whole local landscape (1 km radius)
- Generate all manuscript Tables and Figures (except Figure 1, which was created using external software).
All scripts are written in R.
2. Data/
The Data/ folder contains all raw and processed datasets used to estimate landscape connectivity metrics and quantify contemporary pollen-mediated gene flow. It is organized into three subfolders:
- Genetic Data/
- Hummingbird Data/
- Landscape Data/
A. Genetic Data/
This folder contains estimates pollen-mediated gene flow in Heliconia tortuosa : pollen pool diversity and biparental inbreeding. This also includes summary data for forest patch size, amount of surrounding forest, and pollinator community composition.
- Genetic_Data_Master.csv: This dataset contains information on maternal plants and mating system parameters across different years and populations.
- Mom_ID: Unique identifier for each maternal plant
- Year: Year in which the data were collected
- Pop: Population identifier
- n_Seeds: Number of seeds sampled from the maternal individual
- h_Pollen_Pool: Genetic diversity of the pollen pool contributing to the seeds
- Bootstrap_SE: Standard error associated with the pollen pool estimate
- Outcrossing: Estimated outcrossing rate
- Outcrossing_SE: Standard error of the outcrossing rate
- Biparental_Inbreeding: Estimate of inbreeding due to mating between related individuals
- Biparental_Inbreeding_SE: Standard error of the biparental inbreeding estimate
- Area_ha: Area of the focal forest patch in hectares
- Percent_Forest: Percentage of forest cover within a 1 km radius surrounding the population
- Elevation: Elevation of the site (meters above sea level)
- Proportion_High_Mobility: Proportion of pollinators classified as having high mobility
- SE_High_Mobility: Standard error of the proportion of high-mobility pollinators
- Haplotype_Diversity_Pollen_Sites.csv: This dataset summarizes pollen pool diversity and sampling effort at the site level.
- Site_ID: Unique identifier for each study site
- n_Seeds: Total number of seeds sampled at the site
- h_Pollen_Pool: Estimate of pollen pool diversity at the site
- Boostrap_SE: Standard error of the pollen pool diversity estimate
B. Hummingbird Data/
This folder contains summary data for each focal forest patch.
- Hummingbird_Data.csv: The dataset reports total number of captures per hummingbird species at each forest patch, total number of high-mobility hummingbird captures, proportion of high-mobility hummingbirds per forest patch.
- Patch: Unique identifier for each habitat patch
- Area_ha: Area of the habitat patch in hectares.
- Percent_Forest: Percentage of forest cover within 1 km radius surrounding the patch
- Elevation: Elevation of the site (meters above sea level)
- VISA: Species code for Campylopterus hemileucurus
- GREH: Species code for Phaethornis guy
- LBIH: Species code for Phaethornis longirostris
- CHRH: Species code for Amazilia decora
- GCBR: Species code for Heliodoxa jacula
- RTAH: Species code for Amazilia tzacatl
- SBRH: Species code for Phaeochroa cuvierii
- SNOB: Species code for Amazilia edward
- STHR: Species code for Phaethornis striigularis
- Total_Hummingbirds: Total number of hummingbird individuals
- Total_Hummingbirds_Correction: Corrected total number of hummingbirds
- High_Mobility_Hummingbirds: Number of individuals belonging to species classified as highly mobile
- High_Mobility_Hummingbirds_Correction: Corrected count of high-mobility individuals
- Proportion_High_Mobility: Proportion of the hummingbird community composed of high-mobility species
- Proportion_High_Mobility_Correction: Corrected proportion of high-mobility species
- SE_High_Mobility: Standard error of the proportion of high-mobility hummingbirds
C. Landscape Data/
The dataset includes geographic coordinates of sampled maternal plants, forest patch–level deforestation metrics, and GIS shapefiles and raster layers representing landscape structure.
- Buffer_Area: Contains spatial files representing the full 1000 m buffer area associated with each focal forest patch.
- Buffer_Clipped_Landscapes_Raster: Rasterized versions of each local landscape classified into habitat (mature forest) and non-habitat. Includes a 25 m buffer around each forest patch.
- Clipped_Landscapes: Polygon shapefiles of mature forest patches within each 1000 m buffer landscape. Habitat definition considers mature forest only.
- Clipped_Landscapes_Raster: Rasterized landscapes classified into habitat (mature forest) and non-habitat. Correspond to the 1000 m buffer landscapes.
- Data_Moms_Locations_Mean.csv: Mean spatial coordinates of maternal plants within each focal forest patch.
- Data_Moms_Locations.csv: Spatial coordinates for all sampled maternal plants.
- Data_Patches.csv: Summary data for all sampled focal forest patches. For each focal forest patch, this file includes deforestation variables (amount of forest cover and forest patch size) and data on the local pollinator community composition (proportion of high-mobility hummingbirds).
- Pop: Unique identifier for each population
- n_Seeds: Number of seeds sampled for the population
- Area_ha: Area of the habitat patch in hectares
- Area_ha_QGIS_1km Area calculated within a 1 km buffer around the population
- Percent_Forest: Percentage of forest cover
- Percent_Forest_QGIS_No_Focal: Forest cover percentage excluding the focal patch
- Percent_All_Forest: Percentage of all forest types
- Percent_All_Forest_No_Focal: Percentage of all forest types excluding the focal patch
- Elevation: Elevation of the site (meters above sea level)
- Proportion_High_Mobility: Proportion of pollinators classified as highly mobile
- SE_High_Mobility: Standard error of the high-mobility proportion estimate
- Locations: Contains shapefiles related to maternal plant locations, including Moms_Locations (individual maternal plant points), Moms_Locations_Mean (mean coordinate per patch), and Moms Locations_Median (median coordinate per patch).
- Merged_Forest_Cover_Types_Raster: Rasterized landscapes classified into habitat (all forest types combined) and non-habitat. Includes all forest cover types.
- Merged_Forest_Cover_Types_Raster_Buffer: Rasterized landscapes classified into habitat (all forest types) and non-habitat. Includes a 25 m buffer around each forest patch.
3. Output/
This folder contains .RData objects and tables used in the statistical analyses and reported in the manuscript. All .RData files are R objects containing processed connectivity metrics derived from landscape data and used in downstream statistical models.
A. RData Objects
patch_based_metrics.RData
Contains: Landscape connectivity metrics (structural and functional) and patch area metrics.
These metrics are restricted to the focal patch in each landscape and represent intra-patch connectivity and patch-level attributes used in statistical analyses.
- Pop: Unique identifier for each focal forest patch
- Area_ha: Area of the focal forest patch in hectares
- Area_ha_QGIS_1km: Area calculated within a 1 km buffer around the focal forest patch
- Si.base: Baseline connectivity metric calculated using default assumptions
whole_landscape_metrics.RData
Contains: Landscape connectivity metrics calculated for the entire local landscape (1 km radius) and PLAND metrics (proportion of landscape composed of habitat).
These metrics represent local landscape-scale connectivity and habitat amount for each local landscape. PLAND = Percentage of landscape in focal habitat (%).
- Pop: Unique identifier for each focal forest patch
- Percent_Forest: Percentage of forest cover within a 1 km radius surrounding each focal forest patch
- Pct.forest: Forest cover metric used in connectivity models, incorporating weighting
- Pct.forest.nowght: Unweighted version of the forest cover metric
- Pct.allforest: Forest cover metric including all forest types in the landscape
- Pct.allforest.nowght: Unweighted version of total forest cover across all forest types
- Si.base: Baseline connectivity metric using default assumptions
- Si.base.nowght: Unweighted baseline connectivity metric
- Si.probcross: Connectivity metric weighted by probability of crossing between habitat patches
- Si.probcross.nowght: Unweighted version of the probability-based connectivity metric
- Si.allforest: Connectivity metric including all forest types in the landscape
- Si.allforest.nowght: Unweighted version of the all-forest connectivity metric
matrix_based_metrics.RData
Contains:
- Landscape connectivity metrics calculated for the landscape outside of the focal patch (non-focal matrix) in each landscape. These metrics were ultimately not included in final statistical analyses due to the results.
B. Tables/
This subfolder contains:
- Output tables used in the manuscript
- Model summaries
- Supporting tables included in supplementary materials
These tables correspond directly to results reported in the main text and supplementary materials.
Code/software
All analyses were conducted in R using freely available packages (sp, raster, sf, rgdal, adegenet, vegan, lme4, MuMIn, ggplot2, dplyr, tidyr). The Code/ folder contains all scripts to process data, calculate structural and functional connectivity metrics (including simulated landscapes), run statistical analyses, and generate manuscript figures and tables. All raw and derived data in Data/ and Output/ can be viewed and analyzed entirely with R.
Access information
Some of the data included in this repository were derived from previously published sources:
- Genetic data: Torres-Vanegas F, Hadley AS, Kormann UG, Jones FA, Betts MG, Wagner HH. Tropical deforestation reduces plant mating quality by shifting the functional composition of pollinator communities. J Ecol. 2021; 109: 1730–1746. https://doi.org/10.1111/1365-2745.13594
- Hummingbird capture data: Hadley, A.S., Frey, S.J.K., Robinson, W.D. and Betts, M.G. (2018), Forest fragmentation and loss reduce richness, availability, and specialization in tropical hummingbird communities. Biotropica, 50: 74-83. https://doi.org/10.1111/btp.12487
Both sources are publicly accessible. All other data (landscape spatial data, simulated rasters, derived connectivity metrics) were collected or generated ithe authors for this study.
