Investigating phenological variability of beech forests across Europe using satellite data

Climate change has a major impact on the current environment, with vegetation phenology being the earliest indicator of these effects. Long-term phenological observations, such as those provided by satellite remote sensing, are fundamental for understanding spatio-temporal forest dynamics. Normalized Difference Vegetation Index (NDVI) data represent a well-known proxy for monitoring forest productivity and detecting seasonal variations. The objectives of this work are to identify phenological clusters of beech forests, and to quantify the role of geographic and physiographic variables in the phenological timing of each cluster.  The research focuses also on examining the influence of environmental variables on the mechanisms of phenological response to climate change. To this end, we used the EU-Forest dataset to derive the beech forest location across Europe. Then, for each location, NDVI data were extracted from the Moderate Resolution Imaging Spectroradiometer (MODIS) Terra and Aqua sensors, from 2003 to 2023, with spatial resolution of 250 m and temporal frequency of 8 days. To identify groups of different forest types with similar seasonal timing (i.e., pheno-clusters), we carried out K-means Cluster Analysis on the NDVI temporal profiles. Finally, we characterized each pheno-cluster based on latitude, elevation, temperature, and precipitation, to identify gradients and discriminant environmental conditions. Results showed that the obtained pheno-clusters follow a clear elevation gradient, with a high variability at local scale even within the same macroclimatic conditions. This study indicates that characterizing vegetation phenology can provide valuable information about how forests ecosystems respond to both environmental conditions and climate change.