A DF Structure question: Is it known what the df.world-data.xml:world_region.tree_tile_* fields are used for and the logic for how they're populated?
As far as I can see each biome type present in the region gets one index, but all biomes of the same type in all regions seem to get the same selection of trees in _1 and _2 (with _good, _evil, and _savage have that tree present or not depending on the region's evilness/if that biome within the region has any savage tiles. In both cases the trees actually have to be valid for the region, excluding e.g. mountains and oceans), regardless of whether they're actually present in the selection of plants available in the region or not.
For instance, temperate freshwater marsh gets willow trees in both the _1 and _2 tile despite there being quite a few different trees that grow in such biomes.
I'm both looking for the logic behind how the trees for these fields are selected (to be able to replicate it without simple hard coding, which doesn't fare well in the face of modding or updates), as well as what the fields are used for (if anything: I wouldn't be surprised if they're not actually used).