Last time in our GeoGuide, we explored Digital Elevation Models (DEMs) — the saturated canvas behind many of the maps, analyses, and visual narratives in geospatial work. When rendered thoughtfully, elevation data becomes more than just a background layer; it transforms into a visual story in its own right. In this article, we’ll take a practical look at Relative Elevation Models (REMs) — what they represent, and how you can create them yourself.
Understanding Relative Elevation Models
REM expresses how high or low a point lies relative to a local reference surface, rather than to sea level. Instead of showing absolute altitude, REMs reveal subtle topographic variations that can be hidden in traditional DEMs.
The concept is simple: you remove the large-scale elevation trend (for example, the overall downstream slope of a river) to highlight the microtopography.
In practice, an REM can be derived in several ways:
- Polynomial Detrending (QGIS or Python):Fit a smooth surface (often a low-order polynomial) through the DEM and subtract it:
REM = DEM – f(x, y)
where f(x, y) is the fitted trend surface.
- River-Based Interpolation:
Estimate the river’s surface using sampled channel points and interpolate it (IDW or Kriging). Subtract that interpolated surface from the DEM to show relative heights within the valley:
REM = DEM – RiverSurface
- Normalization by Local Mean:
For large-area studies, compute the mean elevation within a moving window and subtract it to emphasize local relief:
REM = DEM – DEMlocal
Each approach produces a slightly different visual and analytical outcome. For fluvial geomorphology, REMs help delineate paleochannels and levees; for archaeology, they expose faint terrain traces of past human activity.
Making Your First REM
Ready to create scientifically meaningful terrain art? Here’s how to get started:
- Gather Your Data: Download a high-resolution DEM covering your area of interest (e.g.: river)
- Choose Your Method:
- IDW Method (QGIS) — Inverse Distance Weighting for interpolating river surface
- Cross-Section Method (QGIS) — Creating perpendicular profiles
- RiverREM Package (Python) — Automated generation with minimal input
- remr Package (R) — Statistical computing workflow
- Style with Purpose: REMs shine with perceptual color ramps (viridis, batlow) and subtle hillshading that emphasizes relative height patterns
Here we present REM visual based on the River-Based Interpolation with IDW Method performed in QGIS.
Another example, this time from the archaeological side, using Normalization by Local Mean. High-resolution DEMs can reveal subtle topographic patterns that point to forgotten landscapes. By removing large-scale terrain trends or using REMs, we can expose faint traces of ancient roads, buried walls, or settlement mounds, features invisible to the naked eye but preserved in elevation data.
Why Elevation “Maps” Capture Attention
Depth Through Traditional Techniques
With hillshades, hypsometric tints, contours, and relief shading, you can turn a raw DEM grid into a map that practically pops off the page. These classical cartographic methods have been refined over centuries, creating intuitive visual hierarchies that guide the eye across terrain.
The Geospatial Aesthetic
Overlay subtle grids, point clouds, or wireframes, and you lean into that “geospatial aesthetic” — it suggests data, precision, and modeling. This technical visualization style has become increasingly popular in data-driven storytelling, bridging the gap between scientific analysis and visual communication.
The Art-Science Intersection
Maps with “detrended” surfaces (like REMs – Relative Elevation Models) often yield abstract, flowing patterns that verge on art — and yet hold scientific meaning. As cartographer Dan Coe writes:
“Many REMs translate into wonderful flowing artistic images … a visual time capsule of a river’s natural history.”
These visualizations reveal historical river channels, migration patterns, and floodplain dynamics that might otherwise remain hidden in standard elevation data.
Neural Networks Meet Cartography
Recent techniques in neural network-based cartographic shading allow automated elevation visualizations that mimic hand-drawn quality. Research by Jenny et al. (2020) demonstrates that U-Net neural networks trained on Swiss topographic maps can replicate the expressiveness of hand-crafted relief shading in seconds — a process that traditionally required highly specialized cartographers and laborious manual work.
Revealing Hidden Patterns
When combined with historical maps or cultural layers, DEM-based visualizations can reveal subtle spatial patterns (old terraces, river control works, soil lines) that might otherwise be invisible. This multi-layered approach transforms elevation data from a simple terrain representation into a palimpsest of human and natural history.
Further Resources
- UP42 DEM Guide — Comprehensive technical overview of DEMs, DSMs, and DTMs
- ArcGIS Pro Spatial Analyst — Tools and techniques for DEM analysis
- Mapscaping Article — A deep-dive into DEM applications and visualization
Did you like this article? Read more and subscribe to our monthly newsletter!
#Fun
Next article
Elevation data plays a role in many aspects of the modern geospatial world. Regardless of the project, whether it is flood risk modeling, wind farm planning, infrastructure design, or 3D city visualization, a Digital Elevation Model is essential. These datasets serve as a representation of Earth’s surface. Beyond the acronyms DEM, DTM, and DSM, there is a story of technological innovation, precision, and perspective.
Why Elevation Data Matters
Every process that interacts with terrain — from water flow and erosion to urban expansion and renewable energy siting — depends on accurate elevation data. For hydrologists, it defines watersheds and flow direction. For engineers, it helps plan roads and pipelines. For environmental scientists, it reveals how ecosystems respond to landform changes.
In the age of digital twins, self-driving vehicles, and satellite-based climate monitoring, elevation is no longer just a background layer. It’s an active data dimension that influences predictions, decisions, and design.
The History of DEMs
The concept of a Digital Elevation Model (DEM) dates back to the late 1950s, when cartographers at the U.S. Geological Survey (USGS) began digitizing contour maps to calculate terrain heights automatically. Early DEMs were coarse and limited in extent, but the foundation was set: terrain could be represented as numbers.
The real breakthrough came with satellite radar missions. In 2000, NASA and the German Aerospace Center (DLR) launched the Shuttle Radar Topography Mission (SRTM) aboard the Space Shuttle Endeavour, mapping 80% of Earth’s land surface at 30-meter resolution. For the first time, near-global elevation data became freely available — a true revolution for GIS and environmental science.
Over the following decades, missions like ALOS PALSAR (Japan), TanDEM-X (Germany), and Copernicus DEM (Europe) pushed spatial resolutions down to meters or even decimeters, enabling applications that were once unthinkable at continental scales.
The Evolution of Digital Elevation Models (DEMs) by Geoawesome
Understanding the Terminology: DEM, DTM, DSM
Let’s clarify the most common acronyms you’ll encounter:
The DEM is an umbrella term, but the distinction between DTM and DSM is crucial.
A DTM strips away vegetation and structures, exposing the true ground surface. It’s ideal for terrain analysis.
A DSM, on the other hand, includes canopies, buildings, and infrastructure; perfect for energy modeling, urban design, and LiDAR-based 3D visualization.
Other derivatives worth mentioning:
- REM (Relative Elevation Model) – Depicts elevation relative to nearby water bodies, widely used in floodplain and wetland studies.
- nDSM (Normalized DSM) – The difference between DSM and DTM, representing object heights like trees or buildings.
- CHM (Canopy Height Model) – Used in forestry to estimate canopy height from LiDAR data.
- Bathymetric DEMs – Extend below sea level, integrating sonar or LiDAR bathymetry for coastal and marine mapping.
These variations highlight how DEMs are not just “one dataset,” but a family of terrain representations tailored for specific questions about the Earth’s surface.
How DEMs Are Created
DEMs can come from a surprising variety of sources, each with different strengths and trade-offs:
- Photogrammetry – Elevation extracted from overlapping aerial or satellite images. Accurate and cost-effective for local mapping, but limited by vegetation cover.
- LiDAR (Light Detection and Ranging) – Laser-based systems that deliver sub-meter precision, ideal for DTMs and canopy models. Airborne LiDAR, such as that from USGS 3DEP, sets the benchmark for high-resolution elevation in the U.S.
- InSAR (Interferometric Synthetic Aperture Radar) – Uses radar phase differences from satellites (like Sentinel-1 or TanDEM-X) to derive surface elevation and even detect ground deformation over time.
- Radar Altimetry – Measures elevation along satellite tracks, often used over oceans, ice sheets, or remote regions.
Global DEM products like Copernicus DEM (derived from TanDEM-X) or NASADEM (an enhanced version of SRTM) combine radar data and optical imagery for consistent coverage and improved accuracy.
Private companies such as NTT Data (AW3D), Airbus (WorldDEM), and Maxar Technologies (Precision3D) now deliver global elevation data at resolutions down to 0.5 meters, supporting industries from infrastructure to autonomous navigation.
Where to Get Elevation Data
There’s no shortage of DEM sources today — the key is knowing where to look.
Open and free datasets:
- SRTM – 30 m global coverage, foundational for global-scale studies.
- ALOS PALSAR DEM – 12.5–30 m resolution, global.
- Copernicus DEM – 30 m and 90 m resolution, based on TanDEM-X.
- NASADEM – Updated SRTM data with improved void filling and calibration.
- ArcticDEM and REMA – High-resolution polar DEMs from stereo satellite imagery.
Commercial and high-resolution options:
- AW3D (NTT Data) – 0.5–2 m global coverage.
- WorldDEM (Airbus) – 12 m accuracy from TanDEM-X.
- Hexagon Elevation Data – Detailed topographic and bathymetric elevation models.
- Maxar Precision3D – City-scale elevation from stereo satellite imagery.
Data portals and browsers:
- USGS EarthExplorer – Download DEMs by AOI and time range.
- Copernicus Browser – Explore DEMs and Sentinel datasets online.
- Google Earth Engine – Process and visualize elevation data programmatically.
- OpenTopography – Excellent for LiDAR-based and community DEMs, with on-demand processing tools.
- Terrascope – European hub for Sentinel and DEM data analysis in the cloud.
Each of these platforms enables quick visualization, download, and integration of elevation data into GIS software.
How to Visualize and Analyze DEMs
Once you have your DEM, visualization and analysis are where the real insights emerge.
Copernicus DEM, Parco Regionale dei Colli Euganei, Height Map
In QGIS, you can easily:
- Create hillshades for realistic terrain visualization.
Copernicus DEM, Parco Regionale dei Colli Euganei, Hillshade Map
- Derive slope and aspect maps to understand topographic orientation.
Copernicus DEM, Parco Regionale dei Colli Euganei, Slope Map
Copernicus DEM, Parco Regionale dei Colli Euganei, Aspect Map
- Generate contours and profiles to represent terrain structure.
Copernicus DEM, Parco Regionale dei Colli Euganei, Contour Map
Other platforms, ArcGIS Pro or Global Mapper, provide similar functionality with advanced terrain correction, raster algebra, and 3D scene creation.
Elevation in 3D Web Maps
The next frontier for DEMs is interactive 3D visualization. Instead of static maps, users can now explore terrain dynamically in the browser.
- CesiumJS allows streaming and rendering of global terrain datasets, powering digital twins and 3D city platforms.
- Deck.gl integrates elevation data with 3D visual analytics, often used in data storytelling.
- Potree is perfect for point cloud visualization — ideal for LiDAR-derived DSMs.
- Mapbox Terrain and TerriaJS also provide APIs for elevation-driven web mapping.
These frameworks make it possible to build immersive dashboards, from glacier monitoring to infrastructure planning, without specialized desktop software.
The Future of Elevation Data
The next generation of DEMs will be smarter and more dynamic. AI-based enhancement methods are already improving vertical accuracy, noise reduction, and vegetation filtering. Projects like NASA’s Surface Topography and Vegetation (STV) and upcoming TanDEM-L missions promise near-real-time terrain updates.
Meanwhile, open initiatives — such as OpenTopography’s community LiDAR program and ESA’s Open Science initiatives — are democratizing access to high-quality elevation data. Combined with machine learning, these efforts are pushing us toward a future of continuously updated, sub-meter global terrain models.
Every map rests on a foundation of elevation. Whether it’s the slope of a hillside, the rise of a floodplain, or the skyline of a growing city, DEMs define how we perceive and model the physical world. Understanding their differences — and knowing where to find and apply them — is key to unlocking the power of spatial data.
Explore more guides in the GeoGuide Series by Geoawesome and keep building your skills in geospatial fundamentals that move the world forward.
