what is dem data in copernicus

Welcome! I would love to help you unpack exactly what Copernicus DEM data is and how it works.

Before we dive into the technical details, I'd love to know what brings you to Copernicus DEM data today. Understanding your goals helps me tailor our session perfectly!

I want to learn how it compares to other elevation datasets.

That's a fantastic goal. Comparing Copernicus with datasets like SRTM and ASTER is one of the best ways to understand its unique value in GIS and remote sensing.

Your path

Key Resolution Instances (GLO-30, GLO-90)

Data Quality and Editing Enhancements

Common Uses and Applications

What Exactly is a DEM anyway

Before comparing datasets, we need to clear up some vocabulary. In spatial science, the umbrella term Digital Elevation Model (DEM) is often used loosely, but it actually splits into two very different products depending on what the sensors "see" on the ground: a (DTM) and a (DSM).

Side profile of a terrain cross-section comparing a solid green Digital Surface Model line tracing canopy and roof tops against a dashed brown Digital Terrain Model line tracing bare ground.

I've started generating a visualization to help make this difference clear. Notice how the DSM line hugs the tops of the trees and buildings, whereas the DTM line ignores them entirely to stick to the bare earth. This is our first major clue for our comparison: Copernicus DEM is actually a Digital Surface Model (DSM), which means it includes vegetation and buildings rather than bare earth.

Is SRTM a DTM or a DSM?

You've hit on one of the most common points of confusion in satellite geodesy.

The (SRTM) is primarily a Digital Surface Model (DSM). Because it used C-band radar to map the Earth from space back in 2000, the radar signals couldn't fully penetrate dense vegetation. Instead, they bounced off the tree canopies and building tops, capturing a reflective surface rather than the bare ground underneath.

Even though both SRTM and the Copernicus DEM are technically DSMs, they behave quite differently because of when and how they were made. Let's look at how they stack up against each other.

Feature	SRTM	Copernicus DEM
Primary Mission Era	Year 2000	2011 to 2015
Sensing Technology	C-band & X-band Radar	TanDEM-X Radar Mission
Global Resolution	~30m (1 arc-second)	~30m & 90m (GLO-30 / GLO-90)
Vertical Accuracy	Under 16 meters	Under 4 meters

As you can see from the comparison, Copernicus DEM represents a massive generational leap in vertical accuracy, reducing the height error margin from under 16 meters down to under 4 meters! This makes a world of difference when you are modeling water runoff or orthorectifying high-resolution satellite imagery. Would you like to explore how Copernicus DEM achieved this incredible jump in accuracy through the TanDEM-X mission?

Why is vertical accuracy so much better in Copernicus?

That improved vertical accuracy is not a fluke—it is the direct result of a highly sophisticated space mission design.

To understand why Copernicus DEM is so much more precise than SRTM, we have to look at the German radar satellite mission, which gathered the raw data between 2011 and 2015. Instead of relying on a single spacecraft, this mission used a pair of twin satellites flying together in a closely synchronized formation.

Bistatic radar geometry of the TanDEM-X satellite pair in close formation flight.

I've set up a visualization of this unique orbital setup. This dual-satellite structure is called . One satellite acts as the transmitter, sending out a high-frequency radar pulse, but both satellites listen for the echo at the exact same moment. Because the physical gap between them is strictly controlled, we get a highly stable "spatial baseline" that drastically reduces atmospheric distortion and noise. Furthermore, because TanDEM-X used higher-frequency X-band radar, it has a much shorter wavelength than SRTM's C-band, allowing it to capture surface details with razor-sharp resolution.

While this TanDEM-X data forms the raw core of the dataset, Copernicus DEM undergoes significant post-processing and manual editing. It isn't just raw satellite radar; it is an edited derivative of WorldDEM with local infilling to patch up any final voids.

Continue to the different resolution instances (GLO-30 and GLO-90)

Let's zoom in on the different versions of the Copernicus DEM that you can actually download and use.

When working with Copernicus DEM, you'll choose between three main depending on your geographic focus and how much detail your project requires.

Instance	Spatial Resolution	Angular Resolution	Geographic Coverage
EEA-10	~10 meters	0.3 arc-seconds	European countries
GLO-30	~30 meters	1.0 arc-second	Global
GLO-90	~90 meters	3.0 arc-seconds	Global

The global standard for most high-resolution projects is GLO-30. However, there is a catch you need to know when accessing it: GLO-30 is split into a Public tier and a Restricted tier. While GLO-30 Public covers almost the entire landmass of the Earth, a small subset of 30-meter tiles covering sensitive countries has not yet been released to the public. To ensure you don't end up with gaping holes in your maps, standard Copernicus services automatically perform by using the coarser GLO-90 data to patch up those missing public 30-meter tiles.

There's also a fascinating geometric quirk at the poles. Because the Earth is a sphere, lines of longitude get closer together as you move toward the poles. If we kept a constant angular grid spacing of 1 arc-second everywhere, the physical size of our pixels would warp and shrink drastically near the Arctic and Antarctic. To fix this, Copernicus DEM uses a geographic grid adjustment. For areas north of 50°N and south of 50°S, the longitudinal cell spacing is widened to keep the physical ground resolution as close to uniform as possible.

Continue to the editing improvements like water flattening

Let's explore the rigorous editing process that takes raw radar measurements and turns them into a highly reliable, professional-grade surface model.

The Art of DEM Editing

While raw radar signals from the twin satellites provide an incredibly precise baseline, they also capture a lot of natural and physical noise. Ground-level objects, complex hydrology, and sensor anomalies can make the unedited elevation data chaotic. To fix this, the raw WorldDEM dataset undergoes an intensive, systematic editing suite to produce the final Copernicus DEM. One of the most critical steps in this process is water body flattening.

Side-by-side elevation profiles comparing noisy unedited DEM hydrology with corrected smooth river flow and flattened lake surface.

I have created a visualization to show exactly how editing transforms raw hydrology data. On the raw radar side, water bodies look rough and uneven because liquid water absorbs and scatters radar pulses in unpredictable ways. This scatter creates false waves, tilts, and spikes on lake surfaces. To correct this, editors perform on lakes and reservoirs. They identify the boundaries of static water bodies and programmatically set the entire surface to a single, uniform elevation. For rivers, they enforce a rule known as monotonic decrease. Since water only flows downhill, river elevations in the DEM are forced to decrease or stay level as they move downstream. Any artificial bumps, spikes, or dams caused by radar anomalies are smoothed out to ensure realistic drainage modeling.

Beyond inland waters, this editing protocol is applied directly to ocean coastlines to align the sea level to a consistent marine baseline. Spikes, deep pits, and other bizarre artifacts in the terrain are wiped away. Editors also target , such as manually flattening airport runways and ensuring transport infrastructure looks structurally plausible, rather than warped by radar shadows.

What happens if a piece of the globe is completely missing from the radar collection, or obscured by severe radar shadows in deep canyons?