Digital Elevation Model (DEM)

A DEM strictly represents the topographic surface of the bare earth, excluding all above-ground features. This is its primary distinguishing characteristic from a Digital Surface Model (DSM).

• Excluded objects: Vegetation canopy, buildings, vehicles, and other man-made structures are digitally removed.
• Included features: The natural terrain morphology, including ridges, valleys, and slopes, is preserved.
• Hydrological enforcement: Depressions and sinks are often corrected to ensure proper water flow modeling, which also benefits RF diffraction calculations.

A DEM is stored as a regularly spaced grid of square cells, where each cell contains a single elevation value representing the average height for that area.

• Cell resolution: The ground distance represented by one pixel, e.g., 1m, 10m, or 30m. Resolution directly impacts the accuracy of terrain diffraction loss predictions.
• Matrix format: Data is structured as a 2D array of rows and columns, making it computationally efficient for ray-tracing engines to query.
• Georeferencing: The grid is tied to a geographic coordinate system so that every cell corresponds to a precise location on the Earth's surface.

Elevation values are measured relative to a defined vertical datum, which is a reference surface for zero elevation, such as mean sea level.

• Absolute accuracy: The uncertainty of a measured elevation compared to its true value, typically expressed as a Root Mean Square Error (RMSE). High-accuracy LiDAR-derived DEMs can achieve < 10 cm RMSE.
• Relative accuracy: The precision of elevation differences between nearby points, critical for calculating local slope and terrain roughness.
• Integer vs. Float: Elevations may be stored as integers (e.g., centimeters) for compression or floating-point numbers for high precision.

DEMs are generated from various remote sensing technologies, each imparting distinct accuracy and resolution characteristics.

• LiDAR: Airborne laser scanning produces high-density point clouds that are classified to isolate ground returns, yielding the most accurate bare-earth DEMs.
• Photogrammetry: Stereo aerial or satellite imagery is correlated to generate elevation points, but dense vegetation can obscure the true ground surface.
• InSAR: Interferometric Synthetic Aperture Radar from satellite missions like SRTM or TanDEM-X provides global coverage at moderate resolutions (12-30m).

Raw DEM data is computationally processed to extract secondary terrain attributes that directly feed into propagation models.

• Slope and aspect: Calculated from neighboring cell elevations to determine the steepness and orientation of terrain faces.
• Viewshed analysis: Determines which cells are visible from a transmitter location, establishing the Line-of-Sight (LoS) condition.
• Path profile extraction: A cross-section of elevations along the direct path between Tx and Rx, used to calculate Fresnel zone clearance and knife-edge diffraction loss.

For urban RF propagation, the bare-earth DEM is combined with a 3D City Model containing building footprints and heights to create a complete geometric database.

• DEM as base layer: Provides the undulating ground surface upon which buildings are placed.
• Clutter classification: Land-use data (e.g., dense urban, parkland) is overlaid on the DEM to assign nominal clutter heights and loss parameters where explicit building data is unavailable.
• Ray-tracing input: The fused DEM and 3D building model enables deterministic ray-tracing engines to simulate reflections, diffractions, and scattering in complex urban canyons.

The mathematical prediction of radio wave path loss and signal attenuation caused by distance, terrain diffraction, atmospheric absorption, and man-made clutter between a transmitter and receiver. DEMs provide the critical terrain elevation input for calculating diffraction loss over obstacles using knife-edge models. Without an accurate DEM, propagation predictions in non-line-of-sight conditions become unreliable.

A deterministic computational propagation model that simulates the multipath trajectories of radio waves by calculating reflections, diffractions, and scattering from a 3D geometric database. A DEM supplies the terrain surface geometry, while a 3D City Model adds building footprints and heights. Ray tracing engines use DEM raster cells to determine precise reflection points and diffraction edges along each ray path.

A general-purpose, terrain-sensitive radio propagation model that predicts median transmission loss based on irregular terrain morphology, atmospheric refractivity, and surface conductivity. Also known as the Irregular Terrain Model (ITM), it ingests DEM-derived path profiles to calculate diffraction loss over terrain obstacles. Widely used in spectrum management tools for frequencies between 20 MHz and 20 GHz.

A detailed digital representation of urban geometry, including building footprints, heights, and material properties. When combined with a DEM, it forms a complete geometric database for ray-tracing propagation engines. The DEM provides the underlying terrain surface, while the 3D city model adds vertical structures that cause reflection, diffraction, and shadowing in dense urban small-cell deployments.

A spatial layer within a Radio Environment Map that models large-scale, log-normal signal variation caused by macroscopic obstructions between transmitter and receiver. DEM data is used to identify terrain features—hills, ridges, valleys—that create shadow zones. This layer is distinct from distance-dependent path loss and is critical for predicting coverage holes in spectrum opportunity maps.

A discrete global grid system developed by Uber that partitions the Earth's surface into hierarchical hexagonal cells. Provides a standardized, distortion-minimizing spatial indexing system for aggregating DEM-derived terrain metrics and querying REM data. Each hexagon can store average elevation, terrain roughness, and clutter classification for efficient propagation lookup tables.