Digital Elevation Model

A digital elevation model (DEM) is a three-dimensional numerical representation of surface elevation, widely used in geospatial analysis, environmental modelling and planetary science. DEMs provide elevation values for a continuous surface, often as a regular grid or a triangulated irregular network (TIN). Because they capture the height of the terrain or overlying features, DEMs underpin relief mapping, terrain visualisation and a broad range of analytical applications in geographic information systems (GIS).
Closely related terminology includes the digital terrain model (DTM), which represents the bare ground surface, and the digital surface model (DSM), which includes above-ground objects such as buildings and vegetation. Usage of these terms varies in scientific literature and among data providers, although DEM is often employed as a generic label for both DSMs and DTMs.

Terminology and Definitions

There is no universal agreement on the definitions of DEM, DTM and DSM. In many contexts:

• A DSM represents the Earth’s surface including all objects above ground, such as tree canopies and built structures.
• A DTM depicts the underlying bare earth after filtering out vegetation and constructions through processes such as bare-earth extraction.
• A DEM is frequently used as a general term encompassing both DSMs and DTMs.

Some interpretations treat a DEM as equivalent to a DSM, while others define a DEM as a rectangular grid of elevations and a DTM as a more detailed three-dimensional model based on TIN structures. Data sources such as SRTM and ASTER typically provide DSMs; in forested regions these may record heights within the canopy rather than the true ground surface.

Types of DEM Representation

DEMs are commonly represented in two formats:

• Raster DEMs, also known as heightmaps, where elevation values are stored in a regular matrix. Raster DEMs are typically produced through interpolation of remote sensing or survey data and are suited to continuous modelling tasks.
• TIN DEMs, which use irregularly spaced nodes connected by triangles, providing a primary representation of measured points. TINs preserve important terrain features and are efficient in areas with varying data density.

DEMs may be generated through several measurement techniques including photogrammetry, lidar, interferometric synthetic aperture radar (InSAR), IFSAR, stereoscopic imagery and traditional land surveying.

Rendering and Visualisation

Although DEMs consist of numerical elevation values, they are often visualised to enhance interpretation. Common methods include:

• Contour mapping, translating elevation values into lines of equal height.
• Shaded relief, simulating illumination to highlight terrain shape.
• False-colour elevation maps, assigning colours according to elevation gradients, such as transitioning from green at low levels to red and white at high elevations.
• Oblique or perspective views, generating synthetic terrain images.
• Vertical exaggeration, used to emphasise subtle height variations, though this may be considered misleading by some researchers.

These visualisations support tasks ranging from geological assessment to cartographic presentation.

Production Methods

DEM production predominantly involves remote sensing. Traditional methods include manual mapping and interpolation of contour data, still valuable in mountainous regions where interferometry may be limited. Modern DEM creation utilises:

• Interferometric SAR, where two radar passes, or a single pass with two antennas, generate elevation data over large regions.
• Stereo photogrammetry, using paired images taken from different viewing angles.
• Block adjustment and multi-view stereo, applied to aerial photography and optical satellite imagery.
• Lidar altimetry, particularly effective for fine-resolution DTMs.
• Ground-based surveys, including GPS and total station measurements.
• Unmanned aerial vehicle (UAV) surveys, providing high-resolution coverage for smaller areas.

Satellite and Planetary Mapping

Satellite missions have contributed significantly to global and regional DEM datasets. Examples include:

• SPOT and ERS satellites, which provided early large-scale stereoscopic elevation data.
• SRTM, a single-pass SAR mission producing near-global DEM coverage at moderate resolution.
• ASTER, offering elevation data from optical stereo imaging.
• TerraSAR-X and Cosmo-SkyMed, using interferometry for high-resolution mapping.

In planetary science, orbital altimeters have produced detailed DEMs of celestial bodies:

• MOLA for Mars, yielding high-precision global elevation data.
• LOLA for the Moon and LALT for lunar exploration.
• MLA for Mercury.
• New Horizons stereo imaging for Pluto and Arrokoth.

Each planetary DEM is referenced to a defined datum appropriate to that celestial body.

Accuracy and Quality Assessment

DEM quality is assessed using:

• Absolute accuracy, indicating how closely elevations represent true height.
• Relative accuracy, measuring the fidelity of terrain shape and morphology.
• Sampling density, acquisition method, grid resolution and algorithmic processing””, all influencing quality.
• Quality masks, included with some datasets, identifying features such as coastlines, cloud cover, snow and data voids.

DEM comparison against independent datasets is common in quality control.

Applications

DEMs support a wide array of scientific, environmental and engineering tasks:

• Terrain analysis for geomorphology, including slope, aspect and curvature studies.
• Hydrological modelling, simulating water flow, drainage networks, flood risk and mass movements such as landslides and avalanches.
• Soil wetness modelling, including depth-to-water indexes.
• Cartographic applications, producing relief maps and terrain visualisations.
• Image rectification, correcting aerial and satellite imagery for topographic effects.
• Gravimetric corrections, adjusting gravity measurements for terrain influence.
• Infrastructure planning, including line-of-sight analysis for telecommunications and navigation systems.
• Forestry management, assessing canopy height and biomass.
• 3D physical models, including raised relief maps and 3D-printed landscapes.