Sharrukin's Palace

A few years ago, I did a bunch of worldbuilding for an Iron Age fantasy setting that I call “the Great Lands.” I even wrote a full-length novel set there, which got self-published and is currently still available in ebook format on Amazon: The Curse of Steel.

The novel got almost no engagement in its time, and the sequel I had started ran into a plot block, so I eventually set the whole project aside. Lately, however, I’ve been thinking about reviving it – reworking the worldbuilding, pulling the novel down and doing a minor rewrite before republishing it on Royal Road or a similar outlet, and so on.

An early step in the worldbuilding part of the project was to build a new “historical atlas” to settle the back-story of the setting. Which meant, among other things, revising the old master world map. At which point I ran into a snag.

The problem is, I’m too picky when it comes to my worldbuilding. I have to be able to believe in the world, which means I have to pay attention to the earth-sciences part of it, even if none of that is ever going to be too obvious in the finished stories. I know the techniques for developing a constructed world, starting with the plate tectonics and working my way up through the landforms and climate. Yet I’ve never been satisfied with the results when I do that. The worlds I build end up looking too . . . too bland.

I’ve also experimented with random planetary generators – there are plenty of those available, if you know where to look – but none of those come up to my standards. They always end up being too coarse-grained in their details, or if they’re fine-grained they give us naturalistic-looking worlds that make no sense if you examine them more closely. Nobody seems up to the challenge of simulating plate tectonics for a randomly-generated Earthlike world with any degree of fidelity.

So for the past few weeks I’ve been tinkering with worlds, sometimes getting to the point of a world map to start with before throwing the whole thing out, sometimes not even getting that far. Until I had An Idea: if one wants to build a world that’s much like Earth, that clearly evolved under the same physical regime, but isn’t actually our Earth . . . one thing we can do is look at Earth of the distant past.

My original design for the Great Lands involved a Europe-like subcontinent where much of the action would take place. Easy enough. To the south of this, a “Sailor’s Sea” that would allow easy travel from west to east, and then another continent where exotic creatures and cultures might dwell. More continents off in the distance, which might or might not ever become significant to the story.

What I realized was that Earth was actually like that once . . . back in the Eocene Era, before the continent of Africa moved a bit further north and started colliding with Europe and Asia. In that time, what would eventually become the Mediterranean Sea still connected freely both with the (narrower) Atlantic Ocean to the west, and to the Tethys Ocean to the east, an ocean which would eventually become the Indian Ocean once India itself finished making its way north to collide with Asia.

Okay, suppose I work with Earth in the Eocene, about 50 million years before the present. Is it possible to build decent world maps of that era?

Turns out we can. Let me briefly describe my workflow, with pointers to where you can lay hands on similar data and tools if you’d like to fiddle with Earth’s deep past in similar fashion.

The primary resource here is the PALEOMAP Project, work done by the prominent geologist Christopher Scotese. The link will take you to a paper he produced in 2018, describing a set of DEMs (Digital Elevation Models) that he and his colleagues have laboriously assembled for the entire planet Earth in different past eras. This is a monumental data set, with over 110 different maps stretching over half a billion years into the past. There’s a link in the paper that will give you access to the entire data set.

The PALEOMAP models are in a specific file format (NetCDF) that’s in common use in the earth-sciences community, but which I needed specific tools to work with. There’s a NetCDF viewer called Panoply that’s very good for reading the individual files in the PALEOMAP corpus and visualizing the results, but by itself that wasn’t fine-grained enough. I needed to convert the NetCDF files into a different file format like GeoTIFF, so it could be processed by professional cartography software like QGIS.

Fortunately, I was able to locate MyGeodata, a online utility that’s designed to convert geolocation data from one format to another. I was able to convert the two PALEOMAP data files I was most interested in to GeoTIFF with no difficulty. It cost me a few dollars – the size of the datasets were above the site’s threshold for free use – but the results were superb.

I was able to load a GeoTIFF file for the Eocene period (50 million years before present) into QGIS, and work with that to develop a nicely colored elevation map of Earth in the appropriate era. Output from that went into Affinity Photo, and a couple of hours later I had the completed “master world map” at the head of this post.

You should be able to see the differences between Eocene Earth and our present day. The Atlantic Ocean is narrower, and none of the continents have quite reached their present-day positions. The Tethys Ocean is still there, full of islands and island chains that will make it a nice “Sailor’s Sea.” The sea level is noticeably higher than in the present – the Eocene was a rather warm period in Earth’s history, with very little permanent glaciation. Even Antarctica doesn’t have much in the way of ice caps yet. I suspect if I run with that, the “Great Lands” (proto-Europe) are going to be subtropical – but that’s okay.

I’ve added one feature that didn’t exist in our own Earth’s past – an island subcontinent in the middle of the Atlantic Ocean, straddling the mid-ocean ridge. Kind of like Iceland, if it had appeared a few million years earlier and was a different shape. That’s going to be the “Sea Kingdom” in the fantasy setting I’m building, the current highest point of human civilization and the source of world-spanning oceanic adventures.

Okay, so I have a starting point for rebuilding “the Great Lands,” and I ought to be able to proceed from there. Chalked up some neat experience with working with geolocation data and professional-caliber cartography tools, too. Fun!