Data from: Detected shifts towards drought-adaptive strategies in the Amazon Forest over the last four decades

This dataset contains plot- and miniplot-level remote sensing and climate time series for Amazon and Nova Xavantina (Brazil) forest sites used in Barbosa et al. (2026), Detected shifts towards drought-adaptive strategies in the Amazon forest over the last four decades, Global Change Biology. We provide: Raw Landsat spectral bands and vegetation indices at Amazon plots for the Landsat archive period used in the study (see README; broadly covering 1984–2022) (raw_landsat_amazon.csv). Raw Sentinel-2 spectral bands and vegetation indices at Amazon and Nova Xavantina plots for the Sentinel-2 period used in the study (2019–2022) (raw_sentinel2_amazon.csv, raw_sentinel2_nova_xavant.csv). Monthly TerraClimate climate and water-balance variables at Amazon plots for the years used in the analyses (raw_terraclimate_amazon.csv). ECOSTRESS evaporative stress metrics at Amazon plots for the period used in the analyses (raw_ecostress_amazon.csv). Combined Landsat (seasonal coefficients of variation, CV%) for 1984–2022 + TerraClimate (seasonal means) + specific leaf area (SLA) for Amazon plots (landsat_cv_terraclimate_sla_amazon.csv). Combined Sentinel-2 (monthly CV%) for 2019 + ECOSTRESS (monthly means) + SLA for Amazon plots (sentinel2_cv_ecostress_sla_amazon.csv). Combined Sentinel-2 (monthly CV%) for 2019–2022 + SLA for Nova Xavantina miniplots (sentinel2_cv_sla_nova_xavant.csv). All tables include polygon or plot identifiers that can be joined to field and trait data, and correspond to time series derived via Google Earth Engine workflows. The full GEE and R processing code is archived in the companion code repository (Barbosa 2026, Zenodo DOI: 10.5281/zenodo.18268845) and described in the manuscript.