Authors : Subham Roy, Omkar Ghatage, Giribabu Dandabathula
DOI : 10.1007/s41324-025-00612-5
Volume : 33
Issue : 2
Year : 2025
Page No : 1–10
This research evaluates the vertical accuracy of the Forest and Buildings Removed Copernicus Digital Elevation Model (FABDEM), the first open-access bare-earth model with a spatial resolution of 30 m. Using elevation data from the return photons of the ICESat-2 ATLAS sensor as a reference, the study assessed the uncertainties associated with FABDEM in sloped terrains covered with forests and buildings. Three test sites were analyzed: Lavasa, Shimla, and Mussoorie. The evaluation involved calculating the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) and a visual comparison of the 2D elevation profiles from FABDEM and ICESat-2. The error metrics revealed uncertainties of 5.28 m (RMSE) and 4.01 m (MAE) for Lavasa, characterized by low-lying hills. In contrast, Shimla and Mussoorie, located in the Himalayas, showed uncertainties of 8.5 m (RMSE) and 6.5 m (MAE). Nearly 97% of the FABDEM profiles were lower than the canopy heights and buildings, indicating the model's effectiveness in minimizing biases. This research opines that the FABDEM is the best available open-access bare-earth model for these challenging terrains.