Recently I’ve been working on generating world data using a heightmap and various other forms of noise. I start by generating a heightmap that looks like this.
I then use that to create temperature and moisture maps that look like these.
The temperature map combines the height data with the distance from equator along with some other noise to perturb the layers a bit. The moisture map is created by combining multi-fractal noise with the heightmap.
I then take the information generated from these three maps and use a classification system based off of the whittacker diagram seen here. Using this simple classification system you can get fairly good results.
The biome map on the left corresponds to the moisture and temperature maps above. You can see how they affect the formation of biomes. Once this information is calculated I use it to compute a “satellite” image of the terrain. The terrain is generated using color ramps based on the moisture data. Originally thought to interpolate colors based on the biomes but i noticed a striking similarity between the moisture maps I was finding and the colors in satellite images of the earth. I add a few layers of noise to make the map more interesting. here is the final results.
There are things that can be improved. But overall I am pretty happy with the results. I wanted to show off the pictures as soon as I could, I will do a more technical update later to describe some of the methods I used. Especially with the heightmap generation, that was the part I spent the most time on. I also may upload an executable with this generator on it so other people can try it out for themselves!