Global Digital Elevation Model
A seamless, void-filled global digital elevation model — 30 m everywhere, 10 m across many regions, and 1 m across most developed countries.
What this dataset is
A seamless, void-filled global digital elevation model. Resolution is 30 m everywhere as a floor, sharpening to 10 m across many regions where higher-quality public radar or stereo-photogrammetry is available, and to 1 m across most developed countries where national lidar surveys have been released openly. Values are heights above the EGM2008 geoid, in metres.
The blend prioritises the highest-resolution available source for each pixel, falls back through a hierarchy of coarser sources to fill voids, and applies seam-line smoothing across mission and survey boundaries so the multi-resolution mosaic stays visually and hydrologically coherent.
What you get
- A planetary mosaic delivered as 1° × 1° Cloud-Optimized GeoTIFF tiles at the native resolution of each pixel.
- Companion 30 m and 10 m mosaics for users who want uniform-resolution coverage rather than the highest-available blend.
- A void mask indicating which pixels were originally missing and required gap-filling.
- A source-attribution raster indicating which mission or national programme contributed each pixel, and at what native resolution.
Typical uses
- Watershed delineation and hydrological modelling — finer resolution where it matters most.
- Urban flood, surface-water and pluvial-runoff modelling in 1 m countries.
- Visibility, slope and aspect analysis for renewables siting.
- 3-D terrain visualisation in games, simulators and web maps.
- Input to landslide, avalanche and infrastructure-risk models.
Notes on resolution
Resolution varies by region and is documented per-pixel in the source-attribution raster. The 1 m coverage broadly tracks countries with open national lidar programmes (much of Europe, North America, parts of Oceania); 10 m coverage extends across additional regions with high-quality national or regional surveys; 30 m is the global baseline elsewhere.
When working with the highest-resolution blend, be aware that pixel size changes across tile and country boundaries — downstream tooling should read the per-tile resolution rather than assume a fixed grid.
