The idea that you'd have some hard line where desert ends and frozen tundra begins is not appealing to me. So "biomes" will develop naturally where certain terrain intersects certain average climate.
obviously paul is gonna make things as realistic as he can. but i meant how diverse are the biomes gonna be? like theres obviously plains and tundra and desert, maybe a heavy forest area, but what about other stuff? cubeworld got me thinking cuz it has lavalands and undead lands and savannahs, as well as just plain ocean (which is fun since you can sail).
It would be hard to explain or guess as it will be the intersection of a few variables.
As an example, let's say the terrain itself becomes more diverse and spread out... mountains will be higher, sections of the world will be larger, sections of the world will also vary in "roughness". This gives you mountains, mountainous land masses, high plateaus, canyons, large rolling plains, etc..
Water also begins to flow and springs are randomly places. This gives you rivers, streams, lakes, ponds, etc..
Then there is climate and "general weather". A larger fractal will determine the relative average "climate" for an area... basically the min, max, and average temperatures. A particular spot on the world's weather will also be affected by the elevation and the general land layout. (High winds are more common across open flat land or in high elevations, for example.)
It's important to note that all of these are not hard-edged things. Any point on the world will have a blending of these variables.
Finally, we enter the "planting plan". This is a smaller scale fractal that attempts to map the richest soil/growing areas. Basically, it's randomly assigning areas where plants are easier to grow... it will do this irrespective of terrain because a) the terrain will already filter out stuff that won't grow there, and b) it makes it more interesting.
So, for a given block, all of these things combine with the "randomness" to decide what will grow there. If the climate is hot year-around and there are no water sources nearby, then the land is likely to be sand and not support any plant life. If the temperature is cool enough to grow stuff and/or a water source is nearby, then grasses will always be a potential plant unless the "planting plan" is _extremely_ inhospitable for that area. Right here this leads to deserts and grass lands/savannahs that have nice transitions if they happen to be next to each other.
Next up are the trees. A rolling plane with favorable temperatures and extremely high winds will only grow trees in the most hospitable "planting plan" or otherwise near rocky land. Otherwise, you have forests similar to now. If the seasonal temperatures tend to be really cold then the majority of trees will be pines... this would lead to tall pine forests. If the seasonal temperatures tend to be hot and the area is wet then you will end up more rain-foresty feel. Lots of ground cover and high forest canopy.
Areas near shallow standing water will grow marsh plants. I'd like to think they'd be different for full sun salt water than for shaded fresh water... but we'll see. Shallow fresh water can still grow trees if the planting plan is high... then you end up with nice swamps.
The thing is, I don't actually specifically make any biome. I create the environment with the variables above and then when I add support for specific trees and plants, I give them a list of conditions required to grow them. Biomes are a by-product of the fact that all of those other variables will tend to intersect over large areas.
P.S.: I really hope I also find a nice way to add glaciers. The idea of hunting down ice-takwin in pure-ice glacier caves really appeals to me.