Evaluation of GEDI/ICESat-2 Satellite Lidar Datasets for Ground Surface Modelling

Mihnea  CĂȚEANU; Sorina-Mihaela  MICLESCU

Mihnea CĂȚEANU Faculty of Silviculture and Forest Engineering, Transilvania University of Brașov, Romania, România
Sorina-Mihaela MICLESCU Faculty of Geography, University of Bucharest, Romania

Keywords: Lidar; ground surface modelling; geomatics; forestry.

Abstract

This study evaluates the potential of GEDI and ICESat-2 datasets, for modelling the ground surface for mostly forested terrain characterized by complex relief. Both qualitative and quantitative analyses were conducted, using a high-accuracy DTM derived from airborne laser scanning data as reference. Results indicate that ICESat-2 data has higher individual footprint accuracy (RMSE of 1.21 meters, compared to RMSE of 2.66 meters for GEDI). GEDI offers improved spatial coverage and thus higher quality DTMs (RMSEs of 13-14 meters), especially after outlier removal. The research demonstrates that GEDI is more suitable for large-scale terrain modeling due to its significantly better spatial distribution of observations, while ICESat-2 is better suited for estimating ground surface elevation at the footprint level and is ideal for monitoring repeated ground tracks. Overall, we find that both datasets provide valuable insights for ground surface modeling and are unprecedented in terms of spatial coverage and temporal repeatability.