Vegetation phenology along urban-rural gradients across European capital cities

Abstract

Urbanization influences vegetation phenology which feedbacks on climate and ecology. Earlier Start of Season (SOS), later End of Season (EOS) and longer Growing Season Length (GSL) in cities compared to their surroundings is widely observed from optical remote sensing data and is associated with the urban heat island effect. However, these observations are limited by coarse spatial resolution, and finer spatial scale assessments indicate that phenology variation is also driven by vegetation characteristics. In this study we used the new Sentinel-2-based Copernicus High-Resolution Vegetation Phenology and Productivity dataset on a 10-meter spatial resolution to analyze SOS, EOS and GSL dynamics along urban-rural gradients in 37 European capital cities. We analyzed phenology change with linear regression and used random forest to model how land surface temperature, land cover and dominant leaf type influence SOS and EOS in different climate zones and across all cities. We found that urbanization advances SOS and EOS with relatively similar patterns among different climate zones, besides more varied responses in drier climates. Due to the earlier EOS, we did not observe lengthening in GSL, contrary to previous research. Increasing temperatures advance both SOS and EOS, whereas dominant leaf type plays a key role especially in EOS variation with non-tree vegetation having an advancing influence. Our results reveal new insights from urban phenology and its drivers under warming conditions, highlighting the significance of vegetation characteristics and high-resolution remote sensing data for urban phenology analyses.