Similar environments support multiple leaf functional solutions

This archive contains the processed community-level data and R scripts used to derive and analyse multidimensional leaf functional performance in the manuscript. The workflow translates community-weighted leaf traits and environmental conditions into process-based estimates of carbon benefit, water cost, daytime thermal cost and nighttime thermal cost. These derived functional dimensions are then used to compare communities within similar environmental neighborhoods, quantify the dispersion of clear functional advantage, estimate focal slack relative to local efficiency frontiers, and assess the residual dimensionality of leaf functional performance. Files included: sampled_CT.csv: processed global community dataset used as input for deriving leaf functional performance. china_df.csv: processed China transect community dataset used as additional input for deriving leaf functional performance. leaf_data.csv: derived community-level dataset containing standardized environmental variables, community-weighted leaf traits and the estimated functional performance dimensions used in the downstream analyses. 01_calculate_leaf_functional_performance.R: calculates leaf functional performance dimensions, including carbon benefit, water cost and daytime/nighttime thermal costs. 02_analyze_dominance_two_nulls.R: compares communities within local environmental neighborhoods and evaluates whether clear functional advantage is concentrated or dispersed, using two null models. 03_focal_slack_analysis.R: performs focal slack analysis to identify which functional dimensions retain the largest unresolved functional gaps relative to the locally attainable frontier. 04_multivariate_bayesian_model.R: fits multivariate Bayesian response models to estimate residual correlations and effective dimensionality of the functional performance space after accounting for environment and leaf area. The scripts are intended to be run sequentially. The file leaf_data.csv is the main derived dataset for reproducing the dominance, focal slack and multivariate Bayesian analyses. The first script may require access to the environmental raster layers used in the original processing if the user wishes to regenerate leaf_data.csv from the initial input files. All scripts are written in R. Main package dependencies include tidyverse, splash, terra, FNN, lpSolve, mgcv, Benchmarking, brms, posterior, cmdstanr and ggdist.