BlenderGIS Basemap Measurement Accuracy: Web Mercator Distortion Explained
You import a satellite basemap into BlenderGIS, measure a distance you know is 4,000 meters — and Blender says it’s barely 2,800. Nothing is broken. You’re seeing Web Mercator projection distortion, and at higher latitudes it is severe enough to ruin any measurement-based modeling. Here’s how the distortion works, how big the error gets, and what to use instead when accuracy matters.
Quick answer: BlenderGIS basemaps arrive in Web Mercator (EPSG:3857), a projection that stretches east–west distances as you move away from the equator. At 45° latitude, measured ground distances are exaggerated by 1/cos(45°) ≈ 1.414× — so a true 2,828 m stretch reads as 4,000 m. Never use a Web Mercator basemap for measurements or dimension-critical modeling; reproject to a local meter-based CRS (UTM) first.
What Web Mercator Does to Your Measurements
Nearly every web map tile service — the satellite and street basemaps BlenderGIS pulls in through GIS > Web geodata > Basemap — serves tiles in Web Mercator (EPSG:3857). It was designed for fast web mapping, not for measurement. Its defining property: it preserves local shapes and angles (it’s conformal) but inflates distances and areas away from the equator.
The east–west scale factor at latitude φ is 1 / cos(φ):
- Equator (0°): scale factor 1.00 — measurements are correct.
- 30° latitude (e.g., Cairo, New Orleans): factor ≈ 1.15 — a 15% error.
- 45° latitude (e.g., Paris, Portland, Harbin): factor ≈ 1.414 — measured distances are 41% too long. A real 2,828 m reads as 4,000 m.
- 60° latitude (e.g., Oslo, Helsinki, St. Petersburg): factor = 2.0 — distances read twice their true length.
North–south distances are distorted too, by a related (secant-based) formula — the key point is that no direction is trustworthy for measurement at high latitude in Web Mercator. Areas inflate even faster: at 60° latitude, a polygon’s area reads 4× too large.
Where this bites in practice
This isn’t an academic footnote — it bites in very practical ways:
- Architectural context models. If you trace building footprints or measure setbacks off a Web Mercator basemap at 50° latitude, every dimension you derive is ~56% too large. That error propagates into the whole site model. Our ArchiCAD workflow guide and architectural visualization guide both assume correctly scaled context geometry — getting the basemap CRS wrong undermines them.
- Terrain + imagery overlays. A DEM imported in UTM (meters, correct) will not align dimensionally with a basemap draped in Web Mercator — the imagery is stretched relative to the terrain.
- 3D printing and fabrication. Any physical output derived from measured basemap geometry inherits the distortion directly.
Rule of thumb: Web Mercator basemaps are fine for visual context — seeing where things are. They are not acceptable for measurement, dimensioning, or alignment-critical geometry anywhere beyond the tropics. The further from the equator, the worse it gets.
How to Get Accurate Measurements Instead
You have three solid options, depending on your workflow:
Option 1 — Reproject the basemap imagery in QGIS
- In QGIS, load the same area’s imagery (or export the basemap extent) and use Raster > Projections > Warp (Reproject) to convert it from EPSG:3857 to your local UTM zone.
- Import the reprojected GeoTIFF into BlenderGIS via GIS > Import > Import georeferenced raster instead of the Web geodata basemap button.
- Your imagery is now in true meters and aligns with UTM-based terrain and vectors.
Option 2 — Work entirely in a projected CRS from the start
If your project involves terrain, vectors, and imagery together, standardize everything on one meter-based CRS (your UTM zone) in QGIS before importing — the same discipline recommended in our QGIS-to-Blender workflow. Mixed-CRS scenes are where both scale errors and measurement errors breed.
Option 3 — Apply the cosine correction (estimate only)
For a quick sanity check on an existing Web Mercator scene, multiply measured east–west distances by cos(latitude). At 45°: 4,000 m × 0.707 ≈ 2,828 m true. This is an approximation — it corrects the dominant term but doesn’t fix north–south distortion or area — so use it for estimates, never for deliverables.
Verify your scene’s accuracy
Before trusting any measurement in a BlenderGIS scene:
- Check the CRS of every layer. In BlenderGIS, georeferenced imports carry their CRS metadata. Anything in EPSG:3857 is Web Mercator — treat its dimensions as suspect.
- Measure a known distance. Pick a feature with a published dimension (a stadium, a runway, a city block from cadastral data) and compare against Blender’s measure tool. If it’s off by roughly 1/cos(latitude), you’ve found Web Mercator.
- Cross-check terrain against imagery. Import the DEM in UTM and drape the basemap: if the imagery looks stretched relative to the terrain, the basemap is still in Web Mercator.
Frequently Asked Questions
Is Web Mercator distortion a BlenderGIS bug?
No. BlenderGIS faithfully imports the basemap in the CRS the tile service provides — Web Mercator. The distortion is a property of the projection itself, used by virtually all web map services. BlenderGIS can’t “fix” it without reprojecting, which is why the QGIS preprocessing workflow exists.
At what latitude does the error become a real problem?
It depends on your tolerance. At 30° the error is ~15% — already too much for construction dimensions. At 45° it’s ~41%, and at 60° distances double. For any dimension-critical work outside the tropics, reproject out of Web Mercator.
Does the distortion affect north–south measurements too?
Yes, though by a different formula than the simple 1/cos(φ) east–west factor. In practice, treat all directions and all area measurements in Web Mercator as unreliable at mid and high latitudes.
Can I still use the Web geodata basemap button at all?
Absolutely — for visual context and reference it’s fast and convenient. Just don’t trace dimensions off it or align precision geometry to it. Think of it as a picture of the site, not a survey of the site.
What’s the simplest accurate alternative for a small site?
Download the imagery for your site in QGIS, warp it to your UTM zone, and import the GeoTIFF through BlenderGIS’s GDAL raster import. It adds a few minutes and gives you true-meter imagery that matches UTM terrain and vectors exactly.
Does this affect the SRTM elevation import too?
Elevation values themselves are meters regardless of projection, but the horizontal grid of a Web Mercator-sourced DEM inherits the stretching — sample spacing isn’t uniform in true meters. For terrain work, prefer DEMs in a proper projected CRS imported via GDAL.