A Global Extended MODIS-Compatible NDVI Dataset

The study of climate change impacts on vegetation requires access to long-time series vegetation dynamics. MODIS NDVI, with its high chlorophyll sensitivity and data quality, is an important data source in global change monitoring and ecological studies. However, as MODIS NDVI became available only after 2000, the data before 2000 are lacking. This article provides a global MODIS-compatible NDVI dataset at moderate spatial (0.05°) and temporal (16-day) resolution from 1982 to 2000. This dataset generates a long-time series of global NDVI products based on MODIS and AVHRR data using multiple optimization machine learning algorithms. It is designed to synchronously capture complex spatial and temporal correlations of multisource data and account for heterogeneity. Compared with MODIS NDVI, R² of this dataset ranged from 0.79 to 0.95, and the mean absolute error was less than 0.06 in most areas. This dataset addresses the problem of the short period of MODIS NDVI data and provides a new data choice for monitoring global vegetation dynamics and ecological studies.