UN adopts the Equal Earth projection as official - perhaps make that as default for sf objects

[jmcastagnetto]

See the press releases at:

• https://edition.cnn.com/2026/09/04/world/un-vote-new-world-map-intl-scli
• https://news.un.org/en/story/2026/09/1168284
• https://press.un.org/en/2026/ga12779.doc.htm

Having the default representation be the Equal Earth projection could simplify the adoption of this UN resolution.

[teunbrand]

Thanks for the report. The short answer to this is going to be no.

For people finding this issue wondering how to plot this new map standard, here is some code to get you started; Change +lon_0=0 to the central longitude of your data.

library(ggplot2)

world <- rnaturalearth::ne_countries()

ggplot(world, aes(fill = region_un)) +
  geom_sf() +
  coord_sf(crs = "+proj=eqearth +lon_0=0")

^{Created on 2026-09-05 with reprex v2.1.1}

Now for the long answer:
The current default behaviour for geom_sf() is to use the CRS of the object that is provided as data.
Deviating from this behaviour will cause surprises. Not only is this a pain with backward compatibility, but also thoughtful transformations encoded into the data would be ignored.

Very often, the default CRS for data is (or is assumed to be) EPSG:4326, which are longitude / latitude coordinates on a map of the entire Earth. The mathematics of many projections are generally applied from EPSG:4326 as starting point. It is just the de-facto default for how data is stored, because that is just a sensible way to parameterise points on a sphere. Changing user's data to use +proj=eqearth is a bit overreach in my opinion in that it misrepresents the data as the user has delivered it.

Moreover, for places near the antimeridian like New Zealand, the +proj=eqearth +lon_0=0 (default, Greenwich-centered) Equal Earth projection will heavily distort that country. Simply because those places are near the most curved parts of the projection To use that as a default will give a terrible out-of-the-box experience for datasets localised there.

None of these reasons deny that +proj=eqearth is better than +proj=merc (it genuinely is if you're doing anything else than navigating ships), but it is impractical to use +proj=eqearth as some default.

[edzer]

The issue is that with unprojected data (EPSG:4326, or more helpful OGC:CRS84) when we plot we implicitly apply a projection (as the plot is flat, unlike the Earth), which is equivalent to +proj=eqc (equidistant cylindrical), with a lat_ts=xxx latitude of true scale chosen as halfway the latitude range. This is essentially Plate Caree for global maps, and a much worse idea than equal area. For any somewhat larger region, equal area is a better idea; one can pick the +lon_0=xxx as halfway the longitude range. Package tmap uses more sensible defaults @mtennekes @Nowosad than sf::plot() (and geom_sf())

[rsbivand]

The EPSG code for WGS 84 / Equal Earth Greenwich is EPSG:8857, for WGS 84 / Equal Earth Americas is EPSG:8858 and for WGS 84 / Equal Earth Asia-Pacific is EPSG:8859.

[mdsumner]

Obligatory nod to West Wing S02E16 (YouTube video by Aaron Sorkin 😁