Development of a global 1 km phenology dataset (1982–2018) utilizing the spatiotemporal fusion of MODIS and AVHRR data

As climate change continues to reshape global ecosystems, accurately monitoring and understanding shifts in land surface phenology (LSP) has become increasingly critical for detecting vegetation responses, assessing ecosystem resilience, and improving predictions of climate–biosphere interactions. While several moderate-to-fine spatial resolution global LSP datasets exist, most are constrained to post−2000 records due to the availability of MODIS era observations, leaving a significant gap in long-term phenological information and limiting our ability to investigate multi decadal trends. To address this, we developed the first global 1 km resolution LSP dataset spanning 1982–2018. By integrating MODIS and AVHRR NDVI data through the enhanced Flexible Spatiotemporal Data Fusion (cuFSDAF) model, we reconstructed pre−2000 MODIS-like NDVI time series with improved temporal consistency and reduced cross sensor bias. Using logistic functions and third-order derivatives, we extracted four key phenological metrics, including the start of the growing season (SOS), maturity, senescence, and the end of the growing season (EOS), enabling detailed characterization of vegetation development stages across diverse biomes. In the Northern Hemisphere (NH), SOS, maturity, and EOS advanced significantly from 1982 to 2018, while senescence was delayed, indicating a widespread lengthening of the growing season. In contrast, the Southern Hemisphere (SH) experienced delayed SOS, maturity, and senescence, with EOS occurring earlier, suggesting a shortening of the growing season and revealing contrasting hemispheric sensitivities to climate forcing. These metrics were validated against ground-based observations from multiple networks, demonstrating strong consistency, high accuracy, and reliable reconstruction of long term phenological signals. As the first dataset of its kind at 1 km resolution over such an extended period, it provides an invaluable resource for studying global vegetation dynamics and serves as a crucial tool for ecological monitoring, climate change research, and terrestrial ecosystem modeling.