FathomDEM+

FathomDEM+ is a global digital terrain model (DTM) from Fathom, the Bristol-based flood-risk modeller. Global satellite elevation models record the tops of trees and buildings and carry instrument errors, while bare-earth LiDAR and IfSAR surveys cover only part of the world because they are expensive and slow to collect.

FathomDEM+ starts from the Copernicus DEM (GLO-30), a surface model produced from TanDEM-X radar data. A hybrid vision transformer and convolutional neural network, trained where LiDAR exists, predicts corrections that remove trees, buildings and detected height biases; its predictors include canopy height, tree cover, building height and Sentinel-1 radar backscatter. The resulting base model, FathomDEM, is then blended with Fathom's compiled LiDAR and IfSAR collection, which covers more than 10 million km² across 29 countries, with geoid and datum corrections applied.

The dataset covers the whole world except Antarctica at ~30 m resolution, and ~10 m in enhanced areas, with metadata on native resolution, date of capture, source data type and certainty rank. The method is published in Environmental Research Letters (2025), and an independent benchmarking study by Guth et al. (2025) in Remote Sensing named FathomDEM and Copernicus DEM the best global DEMs available. Buyers can access it through the Fathom API, the Fathom Portal, GeoTIFF files for on-premise hosting, or partners, for uses including flood and hydraulic modelling, infrastructure planning and hazard mitigation, (re)insurance pricing models, disaster risk and emergency modelling, and digital twins and GIS platforms.