CANOP 2023 - 2026

Coordination : Karine ADELINE (Office National d'Etudes et de Recherches Aerospatiales)

Abstract : With the increased impact of climate change and anthropogenic activities, one of the most critical vegetation ecosystem service is food supply in a context of worldwide population increase. The difficulty is how to combine a substantial food production, sustainable practices and a viable economy while remaining accessible to all. In orchards, agroecology aims at designing production systems relying on the functionalities offered by ecosystems by minimizing environmental pressures and preserving natural resources. Its reach can be broaden with the use of remote sensing data which brings a solution that is non-destructive, high-throughput and dynamic to assess vegetation condition on a large scale. By combining remote sensing and agroecology, the CANOP project wants to give innovative insights, first for tree health characterization against pests and diseases under contrasted managements (phytosanitary products reduction, controlled irrigation) and for different genotypes for a given management (resilient varieties breeding), and second for the optimization of an adapted use of fertilization. Actually, these tree-oriented applications require an appropriate observation scale. While most studies focus on intra- and inter-species variability, the CANOP project targets the leaf scale and the intra-individual variability within a tree canopy. From optical properties measured in the 0,4-2,5 µm spectral range, leaf biochemical traits can be retrieved such as pigments, water and dry matter content. They witness complex physiological processes such as photosynthesis, transpiration, nutrient allocation, growth rate and decomposition. At the tree scale, the objective of CANOP is to map this variability in leaf pigmentation, water stress and biomass. However, four major challenges arise when tackling the centimetric spatial resolution: (1) the anisotropic behavior of leaf optical properties, so far neglected for higher spatial resolutions, (2) the impact of the tree 3D geometry, dominated by leaf distribution and orientation which emphases multiple scattering effects, (3) the spatial upscaling from leaf to the tree and then the orchard from different remote sensing data (including satellite imagery), and (4) the difficulty to relate the biochemical traits with health/nutrition status values (qualitative/quantitative). To address these issues, the ambitiousness of the CANOP project is to combine active and passive optical remote sensing technologies (3D-LiDAR and imaging spectroscopy) with polarization from laboratory measurements and unmanned aerial vehicle acquisitions. Therefore, a large part of the work is dedicated to experiments, but also to data modelling between optical properties and leaf traits, and leaf traits and health/nutrition status. This includes methods from AI data-driven approaches and physics-based ones from the use of radiative transfer models, as well as new models development. Multi-scale robust methods are the key point for delivering tree health/nutrition maps with high estimation accuracies. The studied sites are apricot and peach orchards (3rd and 4th national rank for fruit production) managed by INRAE, respectively at UERI Gotheron near Valence and at Avignon as part of a national orchard network. Finally, the CANOP project is in line with the 4th program “Investing for the Future” (PIA4) funded by the government with two priority research program and equipment (PEPR), “Agroecology and digital” and “Genetics and varietal selection”, respectively co-led and led by INRAE. Thus, the CANOP results and products will have industrial impacts (ex: design of new close-range remote sensing platforms for variety breeding), agri-food societal and economic impacts (optimizing orchard management for better fruit growth and quality by reducing pesticides) thanks to a complete understanding of vegetation functioning with the powerful alliance of remote sensing, modelling and agroecology.