ICOS Talks: Remote sensing of the land biosphere to support quantification of carbon emissions and removals
Time: Tuesday 26th May 2026, 14:00-15:00 CEST.
Agenda:
- Introduction by Rona Thompson (NILU, Norway)
- Philippe Ciais (LSCE): Mapping biomass change in Europe from AI and satellite data
- Marko Scholze (Lund University): Terrestrial Carbon Community Assimilation System
- Discussion and wrap-up
About the speakers
Philippe Ciais is a researcher working at Laboratoire des Sciences du Climat et de l’Environnement in France near Paris. He is an expert of the global carbon cycle and greenhouse gases emissions monitoring. He is the author of over a thousand of peer-reviewed publications including many in Nature, Science and high-impact journals. He was ranked best scientist in environmental sciences by Research.com. Past co-chair of the Global Carbon Project and lead of the carbon cycle chapter in the 5th IPCC Report. He co-leads the Carbon Monitor initiative that provides daily near real time CO2 emissions. Philippe Ciais is an elected member of the French National Academy of Sciences and an elected foreign member of the Chinese Academy of Science.
Marko Scholze (Lund University, Sweden), is a senior lecturer at the Department of Earth and Environmental Science and the AVENGERS project coordinator. His research focuses on quantifying CO2 and CH4 fluxes using inverse modelling and data assimilation approaches.
Abstracts
Philippe Ciais: Mapping biomass change in Europe from AI and satellite data
Mapping aboveground biomass with artificial intelligence enables high-resolution monitoring of forest carbon stocks and their dynamics. Deep learning models, such as U-Net, integrate multi-source satellite data, including Sentinel-1, Sentinel-2, and spaceborne lidar from GEDI, to estimate canopy height and derive biomass through allometric relationships. These approaches produce annual maps at fine spatial resolution, improving detection of gains, losses, and net carbon changes. Validation against independent datasets shows robust performance, though challenges remain in data quality and transferability. AI-driven biomass mapping supports climate mitigation, ecosystem management, and carbon accounting at regional to global scales. It enhances consistency, scalability, and timely decision-making.
Marko Scholze: Terrestrial Carbon Community Assimilation System
The Terrestrial Carbon Community Assimilation System (TCCAS) is built around the D&B biosphere model, which dynamically simulates carbon and water pools, canopy phenology, photosynthesis and energy balance. A suite of observation operators allows the simulation of solar-induced fluorescence (SIF), fraction of absorbed photosynthetically active radiation (FAPAR), vegetation optical depth (VOD) and surface layer soil moisture (SM). The model is embedded into a variational assimilation system that adjusts initial pool sizes and model parameters to match the observations and provides posterior net and gross carbon fluxes from assimilating remotely sensed FAPAR, SIF, L-VOD and surface layer SM.
