The UN resolution to correct the map, measured

On 4 September 2026 the General Assembly adopted, by 164 votes to one with six abstentions, a resolution encouraging the Equal Earth projection and other equal-area projections where relative size matters. This page does not retell the news. It takes the claim underneath the resolution and measures it.

What the resolution says

Togo led the negotiation on behalf of the African Group, with the backing of the African Union. The text encourages governments, schools, international organisations and technology companies to use the Equal Earth projection and other equal-area projections where relative size matters, and to teach the limitations that every flat map has.

It describes correcting the distortion as an act of "cognitive justice and memorial reparation", and says Mercator-style area exaggeration has contributed to the "symbolic minimization" of Africa. Equal Earth, the projection it names, was published in 2018 by Bojan Savric, Bernhard Jenny and Tom Patterson.

One state voted against: the United States. Six abstained: Estonia, Georgia, Lithuania, Moldova, Serbia and Ukraine.

What it does not say

  • It is not binding. General Assembly resolutions are not, and this one is no exception.
  • It does not ban the Mercator projection, in any words. Marine and aviation charts go on using it, and there is a technical reason for that: Mercator is conformal, so a compass bearing is a straight line on it, which no equal-area projection can offer.
  • It does not say Equal Earth is the correct map. There is no correct flat map - the resolution says so itself, and that is the part it asks to have taught.

The claim, measured

The claim under the resolution is that Mercator's area distortion is large enough to change what people judge about the world. That is measurable. Across the 105 objects this site publishes there are 404 pairs where one covers more ground than the other and Mercator paints the other one the same size or larger. Not drawn slightly small - in the wrong order.

Every row is a pair Mercator puts backwards. The true ratio is the first over the second; the painted ratio is the second over the first. Both are above one, which is what backwards means. Sorted by true ratio, with one row per inflated object - its worst.
Covers more groundPainted larger anywayTrue ratioRatio as Mercator paints it
AfricaGreenland13.71.08
Democratic Republic of the CongoNorway5.911.57
KenyaIceland5.541.02
IndonesiaFinland5.391.01
SudanSweden4.141.08
AfricaCanada3.011.55
South SudanUnited Kingdom2.521.13
ThailandBelarus2.461.07

Every area multiple is integrated over the polygon that is actually drawn, not estimated from `sec^2` at a centroid. That is decisive here: Greenland spans 23 degrees of latitude, the centroid estimate reads 11.9 and the whole-polygon figure 16.5, and one of those reverses Africa while the other does not.

The pair that was argued about all week

Africa covers 13.7 times the ground Greenland does. Mercator paints Greenland at 1.08 times the size of the whole of Africa - not comparable to it, larger than it.

Switch to Equal Earth and the figure returns to 13.7 exactly. An equal-area projection paints in proportion to true area, so it cannot put a pair backwards, and it does not for any pair.

Put the other way round: 14 Greenlands fit in Africa by true ground area, and Mercator paints room for 0.9.

Both panels of this pair, drawn to scale

What equal area costs

No flat map gets everything right, and that is a theorem rather than a position. Mercator keeps angles - its shape ratio is exactly 1.00 at every point, Greenland and Africa included - and pays in area. Equal Earth keeps area and pays in shape: it stretches Greenland to a 3.87 ratio between the long and short axes, and Africa to 1.34.

Which is why the resolution says "where relative size matters" rather than "everywhere". That clause is not diplomatic softening. It is this bill.

None of the five projections on this site is the right one. Each is exact about something and pays for it somewhere else, and every page here prints the bill.

Which projections here the resolution is asking for

2 of the 5 are equal-area, which is the class the resolution names. That is not a list written by hand - the projection registry records which axis each projection holds at 1.00 from its own mathematics, and the equal-area class is the one holding the area axis there.

Under those two the table above is empty. Not empty by luck: empty by definition, because a map drawn in proportion to true area cannot paint a larger place smaller.

Measure it yourself

Every number above was computed by this site, and the tool that computes it is right here. Drag a country across the latitudes and watch the readout move, or morph the map between two projections and watch which places shrink.

What these numbers are not

Areas are spherical integrals of simplified 110m boundaries, on a sphere rather than an ellipsoid. Good enough to compare with and to teach from; not good enough for legal, navigational, surveying or GIS work. That was on this site before the resolution and it does not change now.

The vote counts, the wording and the sponsorship come from the primary material listed below. The figures about the map come from this site's own geometry pipeline. The two are not mixed, and should not be.

Sources

On the vote itself the first entry is the primary record and the rest report on it. On the mathematics of a projection, see that projection's own source list.