Remote sensing and artificial intelligence

The aim is to develop new approaches, in collaboration with users, to enable multi-year monitoring at high spatial resolution of the structure, above-ground biomass and functional composition of tropical forests. The expected accuracy should be sufficient to quantify:

• fine-scale variation in forest composition and functions;
• canopy and carbon losses across the full spectrum of natural and anthropogenic disturbances that characterise forest degradation processes, including selective logging, fire, road and infrastructure construction, and natural mortality events;
• carbon gains from intact forests and regenerating secondary forests, including regeneration.

The approach used is based on the fusion of data from different satellites and artificial intelligence, combined with rigorous validation drawing on Pillar 1 field data, and aims to make significant progress in terms of measured variables, accuracy and attribution, compared with existing estimates of forest cover loss.

Pillar 3 activities are divided into three main tasks:

• Creation of high-resolution maps detailing the composition, structure and biomass of tropical forests, starting with maps of tree heights and forest composition. These maps will then be combined to produce maps showing biomass and its variations. Finally, annual maps of forest degradation can be produced.
• Identifying critical ecosystems for conservation using indicators of forest health and carbon storage capacity in forests that have been damaged or destroyed. These indicators and the data from the previous task will be used to produce high-resolution biomass maps that can be used to quantify carbon emissions and removals at the scale of individual plots and the entire African equatorial forest.
• Respecting the 'FAIR' principles for the use, production and exploitation of data by ensuring that partner countries are included in the research design and by generating open data free of charge.

Publications :