EcoCausality: Global Harmonized Dataset for Detecting and Attributing Biodiversity Change (2000–2023)

This dataset supports the study “EcoCausality: A Causal Machine Learning Framework for Detecting and Attributing Biodiversity Change Across Environmental Gradients.” It integrates globally harmonized biodiversity, climate, land-use, vegetation, and pollution variables (2000–2023) across six continents on a 10 km grid. Each observation includes temperature, precipitation, NDVI, pollution, population density, and biodiversity indices. Data were compiled from GBIF, MODIS, CHELSA, CAMS, GPWv4, and BioTIME under FAIR principles. Provided in CSV format for transparency and reproducibility. Dataset anonymized for peer review and fully open upon article acceptance.