The latest blog post from VITO Remote Sensing – written by ForestPaths member Wanda De Keersmaecker – explains how the project uses earth observation data and machine learning to create high-resolution (10 m) maps of forest structure, including the first prototype maps of Flanders (Belgium).

One of the main ForestPaths’ objectives is to co-design, quantify and evaluate policy pathways to optimise the contribution of forests to climate change mitigation. Comprehensive information about forest structure across Europe is an essential part of developing these pathways. Traditionally, forests are monitored using in-situ measurements such as National Forest Inventories (NFI). While these measurements are highly valuable, they come with certain limitations, such as restricted spatial and temporal coverage. 

In her blog post, Wanda illustrated how by leveraging earth observation data and employing machine learning models, these constraints could be overcome and detailed, high-resolution (10-meter) forest structure maps can be generated.

Read the full blog post here.