Forest cover and matrix contrast jointly structure the taxonomic and functional beta diversity of birds in Madagascar's fragmented

This repository contains the dataset and a complete, self-contained R analysis pipeline for a Madagascar forest-bird beta-diversity study. The studies evaluate the impact of different surrounding matrix habitats on avian communities using a three-matrix design: old-growth forest, rice paddy, and pasture. The data spans both local and landscape scales and focuses on real species occurrence paired with functional AVONET traits. The provided R script evaluates taxonomic and functional beta-diversity for both “forest-generalist” and “forest-associated” species groups.  Files Included bird_community_matrix.csvA presence/absence community matrix consisting of 192 sampling points and 86 real Madagascar bird species. species_traits.csvContains morphological and functional AVONET traits for the 86 species. Trait variables include body mass, hand-wing index, bill width, and diet category. This file also assigns each species to a habitat association (forest-generalist or forest-associated). sampling_sites.csvContains metadata for the 192 sampling sites, including habitat types, matrix designs, and measured local forest cover at multiple radii (from 400m to 2000m). landscape_summary.csv Summarizes the matrix types and landscape-scale metrics for 24 distinct landscapes, including the 2km forest cover measurements. bird_data_long.csv A long-format relational dataset (5,607 rows) linking specific bird observations to their corresponding sampling points, paired sites, and landscapes. bird_beta_analysis_FULL.RThe complete data analysis pipeline. This self-contained script handles the setup, computes pairwise and multi-site beta-diversity, fits beta-regression GLMM models, and generates all primary and supplementary figures and tables (including fourth-corner heatmaps).