Data and code for: Management Reshapes Belowground Productivity Allocation Across Global Resource Gradients

This deposit contains the data and code that reproduce all tables and figures in thesingle-authored manuscript "Management Reshapes Belowground Productivity AllocationAcross Global Resource Gradients" (Ecosystems, Manuscript ID ECO-26-0094), byGianmarco Goycochea Casas (Federal University of Viçosa). The study tests whether human management changes the fraction of total net primaryproduction (NPP) that terrestrial ecosystems allocate belowground, and whether thateffect is comparable in strength to climatic and edaphic controls on total NPP, using456 global terrestrial sites across six harmonized biome classes. Contents:- data/derived/: the processed, analysis-ready site-level dataset (site_analysis_dataset.csv) and a full column dictionary.- code/: the analysis code. reproduce_from_derived_dataset.py rebuilds every table and figure directly from the derived dataset (no source database required); run_results_analysis.py documents the full pipeline from the raw database; plus the R covariate-extraction scripts and a pinned requirements.txt (Python 3.11).- results/: all manuscript tables (1–7), figures (1–4, 600 dpi TIFF), cross-validation predictions, and the results summary. Reproducibility: running "python code/reproduce_from_derived_dataset.py" regeneratesall outputs deterministically (fixed seeds). Raw data: the primary field measurements are NOT redistributed here. They arethird-party data from the global terrestrial net primary production database of Rodalet al. (2025), which must be cited separately by its own DOI. The environmentalcovariates derive from HWSD v2.0 (FAO & IIASA 2023, doi:10.4060/cc3823en), the GlobalAridity Index/ET0 Database v3 (Zomer et al. 2022, doi:10.1038/s41597-022-01493-1),WorldClim 2.1 (Fick & Hijmans 2017, doi:10.1002/joc.5086), and Ackerman et al. (2019,doi:10.1029/2018GB005990). License: code under MIT, derived dataset and results under CC-BY-4.0 (see LICENSE.txt).