CarboST

The CarboST project, funded under the Legge 14 program, aims to provide reliable data on CO2 sink/source strength of ecosystems in South Tyrol. Leveraging a dense network of monitoring stations, the project will develop a remote sensing-driven monitoring system. Key activities include standardizing eddy covariance measurements, processing satellite data, and using these in a simulation model to quantify net ecosystem exchange of CO2 (NEP) and net biome production (NBP) and will provide a web-based visualization of the results for public access.

EURAC, leading Work Packages 3 and 6, will provide remote sensing products and gridded meteorological data, deriving fAPAR from Sentinel-2 (S2) over South Tyrol. The work involves collecting and preprocessing S2 data, estimating fAPAR maps using radiative transfer models and machine learning, and performing cloud-masking and gap-filling. Eurac will validate fAPAR, downscaling MODIS products, and time-series analysis. Data outputs, including raster and vector layers like annual GPP images, will be available through standard web services for external applications or desktop clients. These layers will be stored in the Environmental Data Platform (EDP) of EURAC and will also be visible through the interactive client Geobrowser Map of the province of Bolzano.

Contact person: Laura Stendardi laura.stendardi@eurac.edu

WP3: Remote sensing (Led: Eurac). WP3 focuses on deriving the fraction of Absorbed Photosynthetically Active Radiation (fAPAR) from Sentinel-2 (S2) imagery over South Tyrol. This involves collecting and preprocessing S2 L1C data, applying atmospheric correction using Sen2Cor, and estimating vegetation characteristics using Radiative Transfer Models (RTMs) combined with machine learning. Cloud masking algorithms and gap-filling techniques will address cloudiness, with ground-truth measurements validating the results. MODIS data, spanning 23 years, will be downscaled using high-resolution S2 products and machine learning to improve accuracy in mountainous regions. Uncertainty will be quantified, and consistency between downscaled MODIS and S2 data will be ensured. The model will analyze trends in the MODIS GPP time series. WP3 will also generate environmental factors affecting light use efficiency (LUE), including land surface temperature, air temperature, soil water content, vapor pressure deficit, and solar radiation. Existing data products will be collected, interpolated using geostatistical methods, and refined through downscaling. The Water Stress Coefficient (WSC) will be estimated using precipitation and evapotranspiration data, with solar radiation derived from MSG/SEVIRI data.

WP6: Visualization and outreach (Lead: Eurac).  WP6 aims to publish project data outputs on a public platform for experts and the general public. Data, such as annual GPP images, will be accessible via standard web services and stored on EURAC's Environmental Data Platform (EDP), with compliance to WMS and CSW services. The Geobrowser Map of Bolzano Province will also display these layers. Initial steps involve collecting and preprocessing project outputs from other WPs, primarily WP4. Preprocessing includes format conversion, standardizing spatial reference systems, and optimizing raster file sizes. Metadata will comply with ISO 19115 standards and be available in the EURAC metadata catalogue, also harvested by the GEOSS portal. Data publication will include permission settings and backup systems for security. Post-publication, web tools will facilitate data analysis and project dissemination. The EDP platform's Maps portal will offer interactive layer visualization, with layers also published in the Geobrowser Maps of Bolzano Province using WMS services. This integration enhances accessibility. Additionally, manuscripts detailing WP3 results will be submitted to leading journals like Remote Sensing of Environment or IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, and a comprehensive publication will be prepared for a multi-disciplinary journal like Carbon Balance and Management.

The project was kicked off in June 2024.

Activities to be performed in the year 2025:

WP 3:

• Estimation of fAPAR from Sentinel 2 satellite data and validation;
• Calculation of environmental drivers from meteorological and remote sensing data

WP 6:

• Data processing