Colloques du Collège de France - Collège de France Télécharger l’app
Retour au podcast

Colloque - Gianluca Filippa : Keeping an Eye on Green: Detecting Mountain Grasslands Responses to Climate and Land Use Changes from Imagery

23 min 5 s

Colloques du Collège de France - Collège de France Colloque - Gianluca Filippa : Keeping an Eye on Green: Detecting Mountain Grasslands Responses to Climate and Land Use Changes from Imagery Prêt à écouter
0:00 23 min 5 s

Description de l’épisode

Sandra LavorelBiodiversité et écosystèmesCollège de France2025-2026Vers une écologie fonctionnelle du paysage pour un futur résilient et durableColloque - Gianluca Filippa : Keeping an Eye on Green: Detecting Mountain Grasslands Responses to Climate and Land Use Changes from ImageryGianluca FilippaResearcher, Regional Agency for Environment - Valle d'Aosta, ItalyRésuméRecent satellite imagery has achieved the capability to accurately characterize vegetated surfaces at high temporal and spatial resolution, but standard satellite products often fail to capture the complex, fragmented, and heterogeneous spatial patterns of mountain environments. Using tailored satellite-derived products, we quantify the spatial distribution of grasslands and meadows and assess their year-to-year dynamics in response to processes such as pasture abandonment, shrub encroachment, and glacier retreat. In addition, we are developing dedicated algorithms to detect subtle changes in vegetation structure and biomass, including those caused by mowing and grazing. Together, these efforts enable a more accurate quantification of the impacts of climate and land-use change on mountain grassland resources.Gianluca FilippaMy work is focused on the impact of climate change on mountain ecosystems and includes (1) phenology: remote sensing, proximal sensing, field observation (2) snow & water: snow water equivalent measurement and modelling, hydrological modelling (3) Climate data, climate change scenarios, indices of climate change impact (4) R coding: I'm interested in anything can be done with R, i.e. almost anything.