Drafted with AI and reviewed by Noah.

I built a thing this weekend: Geomorphospace. It’s an interactive map of the theoretical space of river deltas. Not the deltas that exist, though those are in there too. The space of every delta that could exist.

Source is here if you want it: github.com/noahkarsky/geomorphospace.

The idea

This one has been rattling around in my head since a junior-year stratigraphy class. It showed up as one of those half-asleep ideas you get right before you drop off, and it never fully went away.

Two old papers are doing the heavy lifting.

The first is Raup, 1966. He was looking at shell coiling in mollusks and realized you could describe basically any shell with a small handful of parameters. Once you do that, you can plot all possible parameter combinations as a space, and then plot the real shells inside it. Suddenly you can see which shells nature actually built, and, more interestingly, the big empty regions where it never did. That’s a morphospace.

The second is Galloway, 1975. He classified deltas by which force wins: the river pushing sediment out, the waves smearing it sideways, or the tides pulling it in and out. Three forces, so you get a triangle. The Mississippi sits way over in the river corner and looks like a bird’s foot. The Nile is wave-dominated and gets that smooth arc. The Ganges-Brahmaputra is tide-dominated and fans out into a funnel of channels.

Geologists have been drawing that triangle on whiteboards for fifty years. It’s always a static picture with some labeled dots on it. I wanted the picture to move.

What it does

You drag around inside the triangle, or you tune the river / wave / tide balance and the sediment grain size with sliders. As you move, a synthetic delta gets drawn procedurally next to you, morphing between birdsfoot, cuspate, and funnel shapes. You can watch the Mississippi turn into the Nile if you drag the wave energy up.

There are also 50 real deltas plotted in the space. Click any of them and it loads into the instrument so you can see where it sits and what shape falls out.

The whole thing is one index.html file with no dependencies. Open it in a browser and it works. No build step, no npm, nothing to install. I like that a lot more than I probably should.

Honesty about the data

The three numbers behind each delta are relative sediment-flux dominance, and they add up to 1. Eight of the deltas use values quoted straight out of Paniagua-Arroyave and Nienhuis, 2024, which quantified the Galloway triangle properly. The rest are literature-consistent placements based on Galloway, Orton and Reading, and Bhattacharya, and they’re flagged estimated in the source so you can tell them apart.

The delta shape itself is a cartoon with a physics accent. It is not a morphodynamic simulation. If you want real coastal modeling there are people doing serious work on that and this is not it. This is a teaching toy meant to make the classification click.

Why I like it

The part that gets me about morphospaces is the empty regions. When you plot the real things inside the space of possible things, the holes mean something. Either the physics forbids it, or nothing has gotten around to evolving it yet, or we just haven’t found one. Deltas are a smaller and more constrained system than mollusk shells, but the same question applies. Poke around the corners and the edges and see what nature seems to avoid.

Go play with it.