
README

PAPER TITLE
Interspecific variation in shorebird population trends in relation to migration stopover habitat

ABSTRACT 
This repository contains the datasets and R code used to analyse interspecific variation in population trends of migratory shorebirds wintering on the coast of Bangladesh. The study evaluates population trends using four statistical approaches: Negative binomial regression, Quasi-Poisson regression, Least-squares regression of log(count+1) and Least-squares regression of log(bounded-count). Species-level ecological and life-history predictors were used to test which factors explain variation in population trends across species. These predictors include breeding range size, non-breeding range size, migration habitat type, body mass, migration distance, maximum longevity, survival probability and reliance on the Yellow Sea region during migration. The repository also contains data and scripts for additional analyses including phylogenetic comparative models and spatial analyses of citizen-science checklist data. All statistical analyses were conducted in R, and the code included in this repository allows full reproduction of all results presented in the manuscript.

PUBLICATION INFORMATION
Repository DOI https://doi.org/10.5281/zenodo.18463845
Publication date 2026-02-03
License Creative Commons Attribution 4.0 International (CC-BY 4.0)
Copyright Copyright (C) 2026 The Authors.
Full citation will be added after publication.

GENERAL INFORMATION
1. Title of datasets
Population trends and ecological predictors of migratory shorebirds.
2. Author Information
Principal Investigator
Name: to be added Institution: to be added Email: to be added
Additional authors will be listed once the paper is published.

3. Date of data collection
Field data were collected during winter shorebird surveys conducted on Sonadia Island, Bangladesh across multiple survey seasons.
Citizen science checklist data span 2013–2024.

4. Geographic location of data collection
Primary field site Sonadia Island, Bangladesh
Approximate coordinates 21.5° N, 91.2° E
Additional spatial analyses used aggregated citizen-science checklist data from mainland China.

5. Funding information
This research was supported by various partners and conservation organisations, details of which are included in the main paper.

6. License information
Creative Commons Attribution 4.0 International (CC-BY 4.0)
Users are free to share and adapt the material provided appropriate credit is given.

7. DOI for repository
https://doi.org/10.5281/zenodo.18463845

DATA & FILE OVERVIEW
Repository file list
Code
shorebird-trend-migration-habitat.R
Main R script containing all analyses used in the manuscript. This script includes:
- population trend estimation across four modelling approaches - AICc-based model selection for ecological predictors - phylogenetic comparative analyses - analysis of eBird checklist spatial data - generation of manuscript tables and figures

Data
trend_data.csv
Species-level dataset containing ecological and life-history predictor variables used in the comparative analyses.

count_data.csv
Raw count data of shorebirds used to estimate species-level population trends.

tracking_data.csv
Location data from GPS-tagged shorebirds used to visualise migration routes and movement patterns.

ebird_checklist_data.csv
Aggregated dataset of eBird checklists per 1° latitude × longitude grid cell across mainland China between January 2013 and February 2024, downloaded on 1 March 2024 form https://ebird.org/data/download. This dataset is used to examine spatial patterns in observation effort and habitat associations.

phylogenetic_tree.tree
Phylogenetic tree used in the phylogenetic comparative analysis (Phylogenetic Generalised Least Squares models).

Relationship between files
The file shorebird-trend-migration-habitat.R loads the datasets listed above and performs all analyses.
The script produces the following outputs:
- population trend estimates - model selection tables - phylogenetic regression results - spatial visualisations of tracking data - manuscript figures and supplementary tables

Additional related data not included
Additional field survey notes and metadata exist but are not included in this repository because they are not required to reproduce the analyses presented in the manuscript.

Dataset versions
Only one version of the dataset is provided in this repository.

METHODOLOGICAL INFORMATION
Data collection
Shorebird counts were conducted on Sonadia Island, Bangladesh using standardised survey protocols. Observers recorded species presence and abundance during repeated winter surveys.
Population trends were estimated for each species using four modelling approaches:
- Negative binomial regression
- Quasi-Poisson regression
- Least squares regression of log(count + 1)
- Least squares regression of log(bounded-count)
These alternative approaches account for differences in count distributions and sampling variability.

Data processing
Species-level predictor variables were compiled from published literature and global trait databases.
Model selection was conducted using Akaike Information Criterion corrected for small sample size (AICc).
Phylogenetic relationships among species were incorporated using Phylogenetic Generalised Least Squares (PGLS) and robust regression models.

SOFTWARE INFORMATION
All analyses were conducted using R.
Primary packages used include:
dplyr ggplot2 purrr MuMIn MASS nlme ape phytools robustbase boot sf
Additional packages are loaded in the analysis script where required.

DATA-SPECIFIC INFORMATION
trend_data.csv
Number of variables: ~12 Number of rows: species-level dataset (~20 species)
Variables:
species Species identifier code used throughout the analysis.
breeding_range Estimated breeding range size.
nonbreeding_range Estimated non-breeding range size.
habitat_use Habitat use during migration.
body_mass Mean adult body mass.
migration_distance Estimated migration distance.
max_longevity Maximum estiamted longevity.
survival Estimated adult annual survival probability.
yellow_sea_reliance Index representing reliance on the Yellow Sea region during migration.
avg_trend Average population trend across four modelling approaches.
negbin_trend Population trend estimated using Negative binomial regression.
qp_trend Population trend estimated using Quasi-Poisson regression.
ls_trend Population trend estimated using least-squares regression of log(count+1).
bd_trend Population trend estimated using least-squares regression of log(bounded-count).

count_data.csv
Raw count data (2009-2023) used for estimating population trends.
Variables include:
species year site month count

tracking_data.csv
GPS tracking dataset for tagged birds.
Variables include:
bird_id datetime latitude longitude

ebird_checklist_data.csv
Aggregated citizen-science checklist dataset.
Variables include:
checklist_count habitat longitude_bin

phylogenetic_tree.tree
Phylogenetic tree file in Newick format used for comparative analyses.

Missing data
Missing values in datasets are coded as:
NA

Notes
All analyses are reproducible by running the script:
shorebird-trend-migration-habitat.R
The script will reproduce all statistical analyses, tables and figures presented in the manuscript.
