Copernicus DEM vs. LiDAR: assessment of landform accuracy representation at regional scale

Copernicus DEM is the new global dataset with the desired coverage and resolution to become the standard elevation dataset to replace SRTM. SRTM represents the Earth’s terrain at the level of February 2000, with a spatial resolution of 1’ and a feature resolution that can cover medium-scale landform features. The GLO-30 Copernicus DEM is based on the WorldDEM topographic data, which has an original resolution of 0.4”, obtained from TerraSAR-X and TanDEM-X SAR interferometry data. The WorldDEM sources were acquired between December 2010 and January 2015, and every surface was acquired twice. GLO-30 dataset is resampled from 0.4” to 1”. The global validation against ICEsat data revealed a RMSE of 1.68 m, a standard deviation of 1.68 m, and an absolute vertical accuracy linear error at 90% confidence interval of 2.17 m.

In the present approach, LiDAR data at 0.5 and 1 m spatial resolution covering two regions over Eastern and Western Romania were used to evaluate the accuracy in landform representation of the GLO-30 Copernicus DEM. The results show that the deformations due to the RADAR acquisition (shortening and layover) are lower than for the SRTM dataset, but the number of voids is bigger, especially in the mountainous areas. The resolution of the geomorphic landforms is superior to SRTM data, with river channels, gullies, and landslides features being very often recognizable in non-forested areas.

In conclusion, GLO-30 Copernicus DEM outperforms SRTM and can be used as a new source of global elevation data, but care is needed when certain types of landforms are targeted by the analysis, especially in forested areas. The most affected by the inconsistencies due to the RADAR acquisition technology is the hydrological features, especially in mountainous areas with forest cover.