Periodic Reporting for period 1 - TREEMAP (Tree-level biomass maps for Africa)
Periodo di rendicontazione: 2024-06-01 al 2025-11-30
Trees outside forests make up a considerable proportion of the tree cover in drylands, but there is currently a lack of efficient and reliable monitoring systems for trees outside forests. Previous research suggests that most forest and landscape restoration is occurring outside of contiguous forests, particularly in agroforestry systems. This implies that trees on fields, in particular those of smallholder farmers, as well as trees in restoration and plantation areas cannot be effectively and rapidly monitored. The ERC project TOFDRY has developed methods and tools that can fulfil these tasks, but the project outputs are not harmonized and not operational. The TREEMAP project aimed at moving research into practical and needed applications that facilitate the use of tree level maps with end-users, such as governmental authorities or NGOs. In summary, we anticipated that TREEMAP will contribute to the accurate assessment of trees in plantation and restoration areas. This will improve national and project based inventories, reporting, and enable accurate calculation of carbon sinks and credits related to tree plantation and landscape restoration, even for small-holder farmers who possess only few on-farm trees. Potential customers are governments, NGOs, restoration practitioners, and companies investing in the planting of trees.
Beyond its methodological contributions, TREEMAP has demonstrated tangible impact through its application in real-world policy and restoration initiatives. Project results have been used in collaboration with regional and international partners to support monitoring of tree-based landscape restoration efforts under AUDA NEPAD’s AFR100 (https://afr100.org/(si apre in una nuova finestra)) where accurate information on trees outside forests is essential for tracking progress, verifying outcomes, and supporting reporting across diverse agroforestry and restoration contexts in Africa. The specific use for AFR100 was mainly to showcase potentials of TREEMAP, discuss with decision-makers how they could be integrated into their monitoring system, a process which is still ongoing before integration decisions can be made. In collaboration with Planet Labs and AUDA-NEPAD, TREEMAP products have also been applied in pilot sites in Saudi Arabia to support monitoring of large-scale tree planting efforts under the Saudi Green Initiative, which aims to plant 10 billion trees over the coming decades. These use cases illustrate how TREEMAP enables scalable, transparent, and repeatable monitoring of restoration impacts, even in challenging dryland environments. Furthermore, project outcomes have been showcased in high-level decision-making and policy forums, including UNCCD’s COP16, where they contribute to advancing evidence-based discussions on land degradation neutrality, restoration monitoring, and climate mitigation.
Beyond its methodological contributions, TREEMAP has demonstrated tangible impact through its application in real-world policy and restoration initiatives. Project results have been used in collaboration with regional and international partners to support monitoring of tree-based landscape restoration efforts under AUDA NEPAD’s AFR100 (https://afr100.org/(si apre in una nuova finestra)) where accurate information on trees outside forests is essential for tracking progress, verifying outcomes, and supporting reporting across diverse agroforestry and restoration contexts in Africa. The specific use for AFR100 was mainly to showcase potentials of TREEMAP, discuss with decision-makers how they could be integrated into their monitoring system, a process which is still ongoing before integration decisions can be made. In collaboration with Planet Labs and AUDA-NEPAD, TREEMAP products have also been applied in pilot sites in Saudi Arabia to support monitoring of large-scale tree planting efforts under the Saudi Green Initiative, which aims to plant 10 billion trees over the coming decades. These use cases illustrate how TREEMAP enables scalable, transparent, and repeatable monitoring of restoration impacts, even in challenging dryland environments. Furthermore, project outcomes have been showcased in high-level decision-making and policy forums, including UNCCD’s COP16, where they contribute to advancing evidence-based discussions on land degradation neutrality, restoration monitoring, and climate mitigation.
During the technical processing and analysis phase, TREEMAP developed and applied workflows, first, to detect newly planted trees using very-high-resolution SkySat imagery from Planet Labs, and second, to estimate the relative vitality status of individual trees. These methods were implemented across 25 extended pilot sites in Saudi Arabia as a starting point. For each pilot Area of Interest (AOI), the project team at the University of Copenhagen produced and delivered harmonised raster and vector datasets detailing the precise geographic location of detected trees and associated indicators of tree vitality. The approach also demonstrated the potential to distinguish between planting holes with no trees planted yet, and where trees are already planted, enabling temporal tracking of plantation activities and early detection of planting success or failure to suggest targeted early interventions. While the generated datasets are designed to feed into monitoring dashboards and reporting systems, the development of such visualization and reporting tools fell outside the technical scope of the TREEMAP project.
The main achievements and outcomes includeare: (1) A concept note and product pitch that has been elaborated together with Planet Labs on user needs: 2(si apre in una nuova finestra) A demo-code framework on harmonized single tree detection (3(si apre in una nuova finestra) and demo-product for single tree detection in plantation areas based on Skysat (https://rs-cph.projects.earthengine.app/view/skysat(si apre in una nuova finestra)). All is publicly available.
The main achievements and outcomes includeare: (1) A concept note and product pitch that has been elaborated together with Planet Labs on user needs: 2(si apre in una nuova finestra) A demo-code framework on harmonized single tree detection (3(si apre in una nuova finestra) and demo-product for single tree detection in plantation areas based on Skysat (https://rs-cph.projects.earthengine.app/view/skysat(si apre in una nuova finestra)). All is publicly available.
TREEMAP disseminated and distributed demo-products to end-users at the UNCCD’s COP 1630 in Riyadh, Saudi Arabia, and to the Joint Research Center in Italy. TREEMAP partnered with Planet Labs, who was setting up in-country institutional arrangements to enable multi-sector use of the maps. TREEMAP products will generate visibility on and verification of landscape restoration efforts that can attract climate finance.
TREEMAP also delivered open-access technical outputs to support further uptake and scalability of individual tree mapping approaches. In particular, the project released a publicly accessible GitHub repository (https://github.com/dgominski/indivtreemapping_demo(si apre in una nuova finestra)) containing demonstration workflows for individual tree detection and mapping, allowing the public to adapt and extend the methods for planning and monitoring large-scale tree-based restoration initiatives. Then, TREEMAP produced a detailed user guide shared with project collaborators, outlining step-by-step procedures for accessing the repository, preparing training datasets, and applying advanced machine learning models to map individual trees both in restoration and plantation settings and in landscapes with existing tree cover.
As an example, TREEMAP outputs contributed to kick-starting large-scale tree monitoring efforts in the Kingdom of Saudi Arabia under the Saudi Green Initiative. This initiative aims to plant 10 billion trees over the coming decades which is equal to rehabilitating approximately 40 million hectares of land; reduce carbon emissions by 278 million tonnes per year by 2030; and protect 30% of the country’s land and marine areas by 2030. Given the harsh environmental conditions across much of Saudi Arabia and the initial absence of systematic, large-scale monitoring mechanisms, TREEMAP provided a technical foundation for monitoring and verifying tree planting activities implemented by contractors and for tracking progress, for timely targeted interventions where needed.
TREEMAP also delivered open-access technical outputs to support further uptake and scalability of individual tree mapping approaches. In particular, the project released a publicly accessible GitHub repository (https://github.com/dgominski/indivtreemapping_demo(si apre in una nuova finestra)) containing demonstration workflows for individual tree detection and mapping, allowing the public to adapt and extend the methods for planning and monitoring large-scale tree-based restoration initiatives. Then, TREEMAP produced a detailed user guide shared with project collaborators, outlining step-by-step procedures for accessing the repository, preparing training datasets, and applying advanced machine learning models to map individual trees both in restoration and plantation settings and in landscapes with existing tree cover.
As an example, TREEMAP outputs contributed to kick-starting large-scale tree monitoring efforts in the Kingdom of Saudi Arabia under the Saudi Green Initiative. This initiative aims to plant 10 billion trees over the coming decades which is equal to rehabilitating approximately 40 million hectares of land; reduce carbon emissions by 278 million tonnes per year by 2030; and protect 30% of the country’s land and marine areas by 2030. Given the harsh environmental conditions across much of Saudi Arabia and the initial absence of systematic, large-scale monitoring mechanisms, TREEMAP provided a technical foundation for monitoring and verifying tree planting activities implemented by contractors and for tracking progress, for timely targeted interventions where needed.