2026-10-14T14:51:09Z https://www.scidb.cn/oai

10.57760/sciencedb.36775 2026-08-25T09:45:11Z

2026-08-25 Global 1km Resolution 1982-1999 Annual Maximum Composite Normalized Difference Vegetation Index (NDVI) Dataset doi:10.57760/sciencedb.36775 en This dataset was constructed on the Google Earth Engine (GEE) platform, integrating the NASA GIMMS 3GV0 dataset, Terra Climate annual climate data, GMTED 2010 topographic data, and MCD12Q1 stable land use type data as multi-source input variables. The annual maximum Normalized Difference Vegetation Index (NDVI) derived from MODIS MOD13A3 for the period 2001–2013 was used as training ground truth to develop an annual pixel-scale training dataset. A multi-source NDVI reconstruction model was established using the Histogram-based Gradient Boosting Regressor (HGBR), yielding a global 1km resolution annual maximum NDVI product spanning 1982–1999. The dataset is distributed in two compressed archives: the first contains annual 1km resolution vegetation annual maximum NDVI data for the globe (excluding polar regions) from 1982 to 1999, all in TIFF format and named according to the convention ndvi_annual_max_YYYY.tif; the second provides accompanying spatiotemporal change analysis results, including the global vegetation change rate raster Global_NDVI_SenSlope.tif calculated using Sen's slope estimator and the vegetation trend statistical significance test result Global_NDVI_Sig_005.tif at the 0.05 significance level. All data products are in TIFF format with the WGS84 geographic coordinate system. Global; NDVI; Maximum value; 1 km; gradient boosting algorithm Song Lifang Mou Xingyu Zhang Ying Xi Wenyang song rui xiang PUBLIC https://creativecommons.org/licenses/by-nc-nd/4.0/ dataset Science Data Bank