Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I understand what the difference is between the Web Mercator projection and Web Mercator Auxiliary Sphere (WMAS). I also understand that both Google and Esri have adopted this projection as their primary projection for their web-maps which is why we have specialized functions that not only re-project between all projections, but are specific functions for WM, such as webMercatorToGeographic. So ultimately I was wondering why we use the WMAS projection, and the reason it has become a standard in web mapping. Is it a purely a result of two spatial giants moving in that direction or was it just solely because of accuracy reasons?

Additional Links:
Mercator Projection

share|improve this question
Good question! I have always wondered about this too. – Devdatta Tengshe Jul 26 '11 at 3:27
up vote 31 down vote accepted

I'm pretty sure we have Google to thank. Take a look at the original EPSG code WKID for Web Mercator. What does 900913 look like? Helps if you're at least a little l33t.

When Google Maps blew up a few years ago (2005ish), everyone started copying Google. This included Virtual Earth/Bing, Mapquest, Yahoo Maps and eventually Esri. Everyone wanted/needed to be compatible with the most popular web mapping platform. It has been the standard ever since.

Edit: per mkennedy's comment, changed EPSG code to WKID

share|improve this answer
Haha, I never noticed that -- those crazy Googlies. – blah238 Jul 26 '11 at 4:38
I think they prefer "Googlers" ;-) – Derek Swingley Jul 26 '11 at 5:29
I'm voting down, because 900913 is not, and was never, in the EPSG dataset. The first occurrence I've found of 900913 is The original EPSG code was 3785, now replaced with 3857. Say wkid (or just code) instead and I'll vote it up again. Disclaimer: I'm on the subcommittee that maintains the database (although I think I wasn't when 3785 was added!) – mkennedy Jul 26 '11 at 14:30
@mkennedy fixed, thanks for the clarification – Derek Swingley Jul 26 '11 at 14:37
A number like 900913 had to be informally assigned by the open source community so people could "get things done"(tm) and not have to wait several months/year until a committee could agree that 3785 was a bad number and instead be replaced by 3875 which is clearly a much better decision. I am not sure down-voting this answer was fair since it is correct (except for the ESPG vs WKID terminology mistake). Upvoted. – Ragi Yaser Burhum Jul 26 '11 at 14:47

As mentioned, it is originally from Google Maps and its tiling system. However, there are some pretty basic, fundamental advantages* that make the choice of WMAS a good one, and that is why others have followed suit.

The Bing Maps Tile System page is an excellent resource. It outlines some of the advantages of the projection, e.g. that it is conformal and cylindrical. The fact that the area of interest can fit nicely into one square tile at the smallest scale is a nice property when dealing with the tiling system. It also mentions the reason for using a sphere: easier calculations -- which in the early 2000s was an important factor when dealing with JavaScript applications.

*Note that the advantages here are defined in the context of bringing web maps to the masses.

share|improve this answer
+1 Just to clarify the points: "conformal" means the maps will show familiar objects with their correct shapes; "cylindrical" means it's likely (and turns out to be the case) that the projection calculations will be simple. (In a cylindrical projection, the y-value on the map depends only on the latitude.) Even greater simplicity is achieved by using a spherical (rather than the slightly more accurate) ellipsoidal model of the earth. To answer the second part of the question now: obviously these choices are not driven by accuracy requirements. – whuber Jul 26 '11 at 13:59
"the whole earth can fit nicely" ... except for areas near the poles (>85 degrees?). – Kirk Kuykendall Jul 26 '11 at 14:13
@Kirk for maps that are focused on populated places and driving directions, then yes, the whole relevant earth fits nicely. – Derek Swingley Jul 26 '11 at 14:39
@All The "fitting nicely" characterization is a bit of a red herring, because it's going to apply to almost every world projection anybody has ever designed. In fact, the inability of the Mercator to work near the poles its one of its principal shortcomings. – whuber Jul 26 '11 at 16:29
Yeah, shouldn't be an issue unless you're trying to mashup progress from a polar expedition in a classroom. I think polar expeditions provide a great learning opportunity. – Kirk Kuykendall Jul 26 '11 at 16:49

Just a additional 2-bits of thought... the drive for 'Mashups' the spawned off of google made them a defacto-standard, and since they support the majority of the populated world in there mapping we see WMAS become a common baseline. It was not really a decision by the industry as a whole; but a drive to embrace and extend to support the widest audience.

Google did it, Microsoft extended it, ESRI embraced it and now we see a common starting point. Not bad, could there have been better answers, sure, but for the amount that we as the industry have to invest in having a free pretty good global-cached basemap, we really can't complain too much.

share|improve this answer

Just to add a bit, think of the shape of a building footprint. Google Maps (and Bing) commonly show data in cities with building footprints and other items with expected shapes such as property lines.

In a Geographic coordinate system (lat/lon), building footprints that normally have 90 degree corners look very strange and squashed - the angles are not preserved. The coordinate system they chose to use had to be applicable for the vast majority of the populated world, giving a decent appearing view of the map. Area measurement was not the priority - just the appearance of entities in the map. For a whole earth coordinate system, a general Mercator projection works well for appearance due to conformality (angles are preserved). They further simplified the work by assuming that the world is a sphere instead of an ellipse, which makes the math for distance measurement much easier.

So, in short, it's chosen for looks and for easier math.

share|improve this answer
Don't any distortions become small, for any reasonable projection, once we zoom in to street-level scales? – gerrit Jun 3 '14 at 18:28
It is true that non-conformal projections can distort angles, but (1) a geographic coordinate system (lat/lon) is not a projection, and (2) there are other non-mercator conformal projections. – Martin F Jun 3 '14 at 19:16

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.