Scientific Context and Objectives
NASA’s GEDI (Global Ecosystem Dynamics Investigation) mission provides measurements of the vertical structure of forests worldwide, using 25‑m diameter lidar footprints sampled every 60 m along eight tracks per orbit. These data deliver unprecedented 3D information on canopy height, ground elevation, biomass, and structural profiles, but their geolocation still suffers from an uncertainty of about 10 m. This inaccuracy strongly affects validation against airborne or field lidar data, as well as the fusion with optical and radar imagery used to produce forest height, volume, or biomass maps.
The objective of GeoGEDI is to improve the geolocation of GEDI footprints over metropolitan France by applying a dedicated correction method.
Method and Input Data
The project relies on the method developed by
Schleich & al., 2023[1], which improves GEDI footprint geolocation using a high‑resolution Digital Elevation Model (DEM), without requiring raw L1A waveforms. The method uses GEDI Level‑2 products, which include ground elevation, together with a 1‑m resolution reference DEM (IGN DEM) pre‑processed using a sliding‑window smoothing at 1‑m spacing.
This approach offers several major advantages:
- Generalization: reliance on a temporally stable DEM, without requiring strict temporal coincidence with GEDI acquisitions.
- Reduced resource requirements: no need to simulate GEDI waveforms and smaller input files compared to L1A data.
- Spatial scalability: applicability to large areas such as the whole of France, and to other regions where a high‑resolution DEM is available.
The processing code is open‑source and available in the
GeoGEDI GitHub repository, maintained by La TeleScop. It includes the full processing pipeline, DEM preprocessing tools, and examples for running the geolocation correction workflow.
Distributed Products
The dataset includes:
- Improved GEDI footprint coordinates: corrected positions for each lidar shot, accompanied by quality indicators assessing the reliability of the geolocation correction.
- Main Level‑2A thematic variables: ground elevation, canopy height, and relative heights (RH0 to RH100), corresponding to cumulative energy percentiles of the waveform.
Corrected GEDI footprints are distributed as an item by orbit in COPC and Parquet formats, enabling remote queries and extraction over user‑defined areas of interest.
