Comparative analysis of global urban land surface phenology between the MODIS and VIIRS products and extraction methods

The reliability of land surface phenology (LSP) derived from satellite remote sensing is crucial for obtaining accurate estimates of the phenological response of vegetation to future climate change in urban ecosystems. Differences in phenological definition and extraction methodology using remote sensing can generate systemic errors in estimating the phenological temperature sensitivity to predict the biological response of vegetation. Here, we evaluated the start of the season (SOS), the end of the season (EOS), and the growing season length (GSL) between the Terra and Aqua combined Moderate Resolution Imaging Spectroradiometer (MODIS) Land Cover Dynamics (MCD12Q2) and the Suomi National Polar-Orbiting Partnership NASA Visible Infrared Imaging Radiometer Suite (VIIRS) Land Cover Dynamics (VNP22Q2) over 1470 urban clusters worldwide. We found that the availability of valid phenological estimation in urban areas was low across climates, and cities without valid phenological data were mainly distributed in tropical and dry climate zones. The phenological comparison for each urban cluster worldwide revealed significant differences between the two products. For SOS, EOS, and GSL, only a small proportion of urban clusters (6.01%, 3.54%, and 1.50%, respectively) exhibited R2 surpassing 0.5, and approximately 31.63%, 37.37%, and 58.57% showed the RMSD values exceeding two weeks. These results remained consistent and robust across various climates, latitudinal gradients, and valid pixel proportions. We also revealed discrepancies in the inter-annual variability of LSP between the two products. These disparities in phenological metrics directly lead to systematic uncertainties in the measurements of temperature sensitivity. Furthermore, urban phenology disparities primarily arise from the different phenology definitions employed by the two products. Standardizing the definition of phenology metrics extraction from remote sensing data is crucial to the comprehensive understanding of phenological responses to urban environments, which thus provides referencing resources for studying how future environmental changes affect vegetation in natural ecosystems.