A high-resolution canopy height model of the Earth.

Area of Science:

• Ecology
• Remote Sensing
• Geospatial Science

Background:

• Vegetation height variation is crucial for the global carbon cycle, ecosystem function, and biodiversity.
• Accurate, high-resolution vegetation height data is essential for managing terrestrial ecosystems, climate change mitigation, and biodiversity conservation.

Purpose of the Study:

• To create a comprehensive global canopy height map at 10m resolution for the year 2020.
• To develop a model that fuses satellite LiDAR and optical data for accurate canopy height retrieval and uncertainty quantification.

Main Methods:

• Developed a probabilistic deep learning model.
• Fused sparse height data from the Global Ecosystem Dynamics Investigation (GEDI) space-borne LiDAR mission.
• Utilized dense optical satellite images from Sentinel-2 for global coverage.

Main Results:

• Generated a global canopy height map at 10m ground sampling distance for 2020.
• The model improves retrieval of tall canopies, which store significant carbon.
• Only 5% of the global landmass has trees taller than 30m, with 34% of these located in protected areas.

Conclusions:

• The developed canopy height map provides crucial data for ecosystem management and climate change studies.
• The approach enhances the understanding of tall forest distribution and their conservation status.
• This work supports forest conservation efforts and advances climate, carbon, and biodiversity modeling.