Recently, a colleague working on WebGIS overlay analysis asked: Does zoom level 13 on Google Maps correspond to a scale of 1:20,000? The question seems simple, but it actually confuses some fundamental concepts between traditional maps and web maps. Today, I’ll try to clarify this.

No Direct Correspondence

Let’s start with the conclusion: Google Maps uses the Web Mercator projection, so zoom levels and map scales cannot be rigidly matched like they can in ArcGIS layout view.

First, scale varies with latitude. The Mercator projection stretches areas at high latitudes. The same zoom level shows different ground distances at the equator compared to Beijing. Second, scale also depends on screen DPI. Web maps are displayed on monitors. What “1 cm on the screen represents in meters” differs completely between a 96 DPI display and a Retina screen. Finally, tiles are 256×256 pixels each. At zoom level 0, the entire world fits into a single 256×256 tile. Each additional zoom level doubles the pixel count in both dimensions, quadrupling the resolution. This mechanism is not the same concept as the scale denominator used in ArcGIS; forcing a direct conversion yields only an approximation.

In other words, zoom level describes which layer of the tile pyramid you are on, while scale describes the ratio between distance on the map and distance on the ground. They are related but not a fixed constant.

The Most Practical Formula

Although there is no direct one-to-one mapping, you can still make a rough correspondence. Here is the official approach shared by a Google employee:

metersPerPx = 156543.03392 * Math.cos(latLng.lat() * Math.PI / 180) / Math.pow(2, zoom)

In this formula, 156543.03392 is the base ground resolution at zoom level 0 at the equator, in meters per pixel. The calculation is as follows:

Equatorial circumference = 2 × π × 6378137 ≈ 40,075,016.686 m
At zoom level 0, 360° of longitude exactly fills one 256-pixel-wide tile.
156543.03392 = 40,075,016.686 / 256

In practice, for low zoom levels, using the latitude at the map center as an approximation works reasonably well. For large regional extents, however, the error between the center and the edges becomes quite noticeable.

Reference Scales by Zoom Level

Both the ArcGIS documentation and the OpenStreetMap Wiki provide intuitive comparison charts. I have pulled two images from the original articles for reference. Both examples are given at the equator.

ArcGIS Developer Documentation — Zoom Level vs. Scale:

OpenStreetMap Wiki — Zoom Level Illustration:

Conclusion

If your project requires it, you can follow the official documentation or use the formula provided in this article. In reality, however, the scale of Web Mercator (EPSG:3857) is most accurate at the equator. Away from the equator, the scale is always enlarged.

If you have better conversion tips or have encountered pitfalls in real projects, feel free to leave a comment and share.