Earth’s shape is something we struggled for centuries to fully understand. Is neither a sphere, nor an ellipsoid, but a complex geometry called Geoid, impossible to represent with a precise mathematical formula. Things get much more complicated when trying to project this 3D geometry to a 2D plain representation. That’s why we might get confused by tons of different maps where continents seem to have different sizes from one map to another. It all depends on the geometric property we want to preserve when projecting from the almost-sphere to 2D projection: areas, angles and distances.

The most common projection is WGS84, a projection that preserves angles, thus all parallels and meridians are straight lines equally distributed. However, distortions arise in distances and areas as we move away from the equator, especially along parallels.

For this reason, after checking hundreds of existing projections (which can be seen here), I finally chose a projection which is a compromise of all the 3 geometric properties: the Loximuthal projection. Parallels are straight lines equally distributed, while meridians are distorted moving from the central meridian and have extremities on both Earth’s poles. While WGS84 coordinates are expressed in degrees, Loximuthal projection is metric.


All raw data was processed on QGIS using open-access resources:

1. DEM elevation was retrieved from Multi-Error-Removed Improved-Terrain DEM (MERIT) dataset. An improved model of the SRTM DEM, integrating data from other DEM sources for error and noise reduction. This model has been further improved by integrating Copernicus's GLO DEM model and GEBCO bathymetry for the representation of seas and oceans. Bathymetry was added for the largest inland water bodies: Caspian Sea, Great Lakes and Lake Baikal. This dataset will be the source height map for each Minecraft Level. The native resolution, or cell size, of this dataset, is 90m subdivided into many tiles but was downsized to 300m to merge everything in a single file (which is still 16GB large).

2. Land cover data was retrieved by Copernicus Global Land Cover, a comprehensive high-resolution dataset for land cover distribution, integrated with other sources to cover Antarctica and the Artic region and manual improvements for better coastal line definition. This dataset will determine the total amount of landmass on each Minecraft Level. The counting of landmass will include also all inland water bodies, including the aforementioned ones. This is a simplification due to the different altitudes of large lakes compared to sea levels. The native resolution of this one is 100m and was downsized to 300m to match the exact size of the other dataset.


Through land cover data, I’ll try to represent landscapes around the world as accurately as possible using the full potential of Minecraft biomes and environmental features. A set of custom trees was also designed for this project, replicating some common tree species and generic shrubs and bushes.

Earth's radius at the equator is commonly known as being 6378 km long, multiplied by two for the diameter and by Pi we obtain the length of the equator line, for a total of 40075,017 km or 40075017 m, which is essentially Earth's circumference. Since Minecraft works in metres, scales will be expressed in metres. The smallest possible scale to represent Planet Earth follows precisely the geometric properties that characterise it. Since 40075017 is an approximation by an excess of the actual estimated length, the first possible scale is 1:40075016 metres. The output world size is 1x1. Beyond that, only deep oceans will make you isolated. Will you survive under such grim conditions?

Things will get more and more interesting, as we increase the scale by reducing the denominator of the scale. I will create a new MC level every time something interesting happens in this process: the highest altitude in a mountainous area, an increased number of landmass cells, and the increase in the output world size either horizontally or vertically. We will be able to see how continents will take shape, how mountain ranges will start to make their presence clear and how different biomes will distribute around the globe. As the concept of scale may create a bit of confusion, from another perspective, the scale can be considered as the size (in metres) of the output cell (or block). For each level, I will provide some background information on the geographical points on Earth plus a map from NASA’s Blue Marble Next Generation mission, with the position of every land point.


The spawn point for each subsequent level will always be on land at random locations since initially, the world size is irrelevant to make the difference.