Improving photosynthetic phenology detection by incorporating vegetation index with meteorological factors

Accurate detection of vegetation photosynthetic phenology is crucial for understanding terrestrial ecosystem dynamics and improving carbon cycle modeling. However, traditional vegetation indices (VIs) often overestimate the growing season length due to the mismatch between canopy greenness and photosynthesis, as well as uncertainties arising from background noise and saturation effects. To address these limitations, we developed a Modified Phenology Index (MPI) that integrates VIs with meteorological constraints to more reliably capture canopy photosynthetic phenology. Using data from 84 carbon flux sites, we compared phenology derived from five VIs, two solar-induced chlorophyll fluorescence (SIF) products, and five MPI variants against GPP-based phenology using multiple extraction methods. MPIs markedly improved the accuracy of phenological detection relative to VIs, outperforming SIF in identifying the end of the growing season (EOS) and performing comparably for the start of the season (SOS). Across the Northern Hemisphere from 2001 to 2022, NDVI-based MPI revealed widespread advancement in SOS (58.5% of pixels) and slight delays in EOS (51.5%). These results demonstrate MPI’s robustness for large-scale photosynthetic phenology monitoring and its value in climate–ecosystem interaction studies.