Reddit reviews The Science of Fractal Images

The original paper: https://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=7D041E4B27365771B6EB7EAA7022FE7D?doi=10.1.1.21.3719&rep=rep1&type=pdf

I gather it's also published in this book (along with loads of other relevant things): https://www.amazon.com/Science-Fractal-Images-Heinz-Otto-Peitgen/dp/0387966080/

Blending in a secondary heightmap has been a favorite trick of mine for these sorts of things as well. I explain more about the approaches I've taken in the past in another comment on this post. Most of what I was doing was focused on creating discrete, stand-alone continents + surrounding smaller islands, or individual islands / island chains, not whole-world-maps. The idea was that I could build up the world of my games incrementally by focusing each one on a small region of the world and then later decide how those pieces fit together and how many there ought to be. Sadly, my game studio wasn't successful enough to pursue that idea in much depth.

Your code looks awesome! I don't quite have the time to dive in and follow along with what it's doing in depth right now, but it seems like a fair bit of the "magic" is in _filterRelief, yes?

(Random aside -- One thing to keep in mind, if you're planning on using your results in a game: If you want this stuff to be efficient in terms of GPU usage, you'll want to clip your texture to an even power-of-two at the end of synthesis, otherwise GPUs will pad it out to the next power-of-two size on each edge. Lots of wasted VRAM!)