Exploring the discrepancies between ground- and satellite-based autumn phenology: a comparative analysis in European beech forests

Autumn phenology, related to leaf senescence processes, remains a neglected aspect in the study of climate change impacts in deciduous forests. Conventional field monitoring and remote sensing technologies have advanced our understanding of autumn phenological dynamics. However, integrating both approaches may lead to a spatial and temporal mismatch related to different observational scales, introducing uncertainties in the interpretation of satellite-derived phenology. To obtain a comprehensive view of the autumn phenology, this study explored the environmental significance of the mismatch between ground and satellite data, with the following objectives: (i) to quantify the temporal discrepancy between phenology from ground observations (Pan European Phenology database) and satellite-derived time series based on the Moderate Resolution Imaging Spectroradiometer (MODIS) Enhanced Vegetation Index (EVI) in European beech forests; (ii) to assess the influence of environmental factors (temperature, precipitation, latitude, elevation) on the mismatch. Ground observations were matched with autumn phenological metrics extracted from EVI temporal profiles in the period 2003–2015. Statistical comparisons and redundancy analysis were then used to assess their temporal discrepancies and explore environmental influences. The results identified two distinct leaf senescence phases, occurring in the mid-early and the mid-late portions of the autumn phenology. Environmental factors influenced mismatch with temperature and precipitation associated with the mid-late and mid-early portions, respectively. Overall, the observed discrepancies reflect different phenomena characterizing autumn phenology during senescence, each shaped by different environmental factors. Ground observations capture specific moments in leaf senescence, while satellite metrics reflect broader canopy-level dynamics. This study highlights the complexity of autumn phenology and the potential for proper integrating and interpretation of remote sensing and ground observations to improve monitoring of deciduous forests in the context of climate change.