DEM
Digital Elevation Model. A 3D representation of terrain surface created from elevation data, excluding buildings and vegetation. Used for slope analysis, watershed delineation, flood modeling, viewshed analysis, and orthorectification of satellite imagery.
Overview
A Digital Elevation Model (DEM) is a raster dataset that represents the elevation of the Earth's surface as a regular grid of height values. Each pixel stores a single value — typically meters above a reference datum such as the EGM96 geoid or WGS84 ellipsoid — describing the vertical dimension of the terrain. DEMs are foundational datasets in geospatial analysis, underpinning applications from hydrological modeling and slope analysis to viewshed computation and 3D visualization. The term "DEM" is often used generically, but technically it refers specifically to a bare-earth elevation surface with vegetation, buildings, and other objects removed.
How It Works
DEMs are generated from several remote sensing techniques. LiDAR fires laser pulses from aircraft or drones and measures the return time to determine ground elevation with centimeter-level accuracy (typically 5–15 cm vertical precision); its ability to penetrate vegetation canopy via last-return filtering makes it the gold standard for bare-earth DEMs. Stereo photogrammetry uses overlapping optical images captured from satellites or aerial platforms; by computing parallax between stereo pairs, elevation can be derived at 30 cm to 1 m vertical accuracy. Interferometric Synthetic Aperture Radar (InSAR) uses the phase difference between two radar acquisitions to derive surface elevation — SRTM used this technique from the Space Shuttle to produce a near-global DEM in just 11 days.
Post-processing steps for all methods include point cloud filtering, interpolation to a regular grid, void filling, and vertical datum alignment.
Key Facts
- DEM is a bare-earth model; a DSM (Digital Surface Model) includes the tops of buildings, trees, and other features; a DTM (Digital Terrain Model) typically refers to bare earth plus supplementary vector features.
- SRTM (2000) produced a near-global DEM at 30 m resolution with approximately 15 m vertical accuracy using single-pass InSAR from the Space Shuttle Endeavour.
- Copernicus DEM, derived from TanDEM-X radar data, is available at 30 m (GLO-30) and 90 m (GLO-90) globally.
- LiDAR-derived DEMs can achieve 5–15 cm vertical accuracy, while spaceborne DEMs are typically accurate to 2–10 m vertically.
- DEM resolution classes range from ultra-high (sub-meter, LiDAR/drone) through high (1–10 m), medium (10–30 m), to low (90 m+).
Applications
Hydrological Modeling
DEMs are used to delineate watersheds, calculate flow direction and accumulation, model flood extents, and design drainage infrastructure.
Infrastructure and Urban Planning
Road and railway route planning, cut-and-fill volume calculations, line-of-sight analysis for telecommunications towers, and site suitability assessment.
Hazard and Risk Assessment
Landslide susceptibility mapping, coastal inundation modeling, avalanche path prediction, and seismic site response analysis use DEMs to quantify terrain characteristics.
Orthorectification of Satellite Imagery
Removing terrain-induced distortion from satellite and aerial images requires a DEM to model how topographic relief shifts pixel positions.
3D Visualization and Simulation
Flight simulators, virtual globes, augmented reality applications, and gaming engines use DEMs to render realistic terrain surfaces.
Limitations & Considerations
All DEMs are approximations of the real terrain surface. Spaceborne radar-derived DEMs like SRTM are actually Digital Surface Models — they capture the first reflective surface (treetops, building roofs) rather than bare ground, introducing systematic positive bias in vegetated and urban areas. Void areas (from radar shadow or water surfaces) require interpolation and can contain artifacts. Resolution limits the representation of small terrain features: a 30 m DEM cannot capture gullies or embankments smaller than its grid spacing. Vertical accuracy degrades on steep slopes. Temporal currency is another concern — the SRTM DEM reflects the year 2000 surface. LiDAR-derived DEMs offer much higher accuracy but are expensive and only available for limited geographic areas.
History & Background
The concept of representing terrain as a regular grid dates to the 1950s and 1960s, when mapping agencies began digitizing topographic maps. A breakthrough came in February 2000 with the Shuttle Radar Topography Mission (SRTM), which mapped 80% of Earth's land surface in just 11 days — producing the first near-global high-resolution DEM. The ASTER GDEM, released in 2009, extended global coverage using optical stereo pairs. The German TanDEM-X mission (2010–present) produced a global DEM at 12 m resolution, with a reprocessed version becoming freely available through Copernicus as GLO-30 in 2021. Meanwhile, airborne LiDAR technology matured rapidly, enabling sub-meter DEMs for national mapping programs. Today, the trend is toward higher resolution through drone-based LiDAR, satellite-borne LiDAR (ICESat-2, GEDI), and AI-enhanced processing pipelines.
Analyze DEM data with LYRASENSE
Use our agentic notebook environment to work with satellite data and compute indices like DEM — no setup required.