I want to stress that this is merely my opinion, but my honest advice would be: don't.
I used to approach mapping my first worldbuilding project in a similar way back when I started out, and the best lesson I think I personally ever learned was to let go of that kind of thinking. A huge-ass map like that, one that includes absolutely anything and everything on said world, would have required both the kind of machine I could never afford and the kind of patience and perseverance I will never have.
My advice is to chop the project into several maps. You can have a larger scale map showing the general view of the entire world, like the one you have here. Use that merely as a base for the smaller-scale regional maps. You can map out each continent or each country on the continents. It makes things easier for your computer as it doesn't need to handle monstrous file sizes, and saves you the, quite frankly, wasted time trying to find textures that would work on a project this size that simply do not exist.
It also gives you the freedom to play a bit more with the visuals of the maps you create. I know that having a world map / atlas with a completely unified look can be appealing, but I'd invite you to consider that having the separate maps be a little different from each other can also be all kinds of fun. Your world is bound to have different people and cultures inhabiting it, and the different maps can reflect this. It both makes the project interesting to look at, and more fun to make as well, as you get to try different things and play with colours and styles.
I don't know how that will work for you, but for me this kind of approach saved my sanity. I'm much, much happier creating several smaller maps I can actually get finished rather than working on one, big ultimate map that I will never, ever complete. I think the first two years of my mapping hobby were spent on trying to make that big ultimate map, and it was an endless cycle of working for hours and hours and starting from scratch as the approach I was trying that time proved ineffective. It wasn't until I finally scrapped that entire idea that I actually started getting anywhere.
So yeah, absolutely do not take this as gospel truth and feel free to disagree, but this is my two cents on this topic. If you are dead-set on approaching the project like this, I would recommend looking to make it entirely vector based - I think that is probably the only way you can accomplish the scale you are wanting without losing quality as you zoom in.
Reply With Quote
