Physical commonalities are easily gameable. You could divide a city into east-bank/west-bank, or upstream/downstream, or even near-bank/far-bank. Give me enough factors and write me a large enough check, and I can give you whatever kind of partisan results you want.
I do find watershed democracy (legislative district boundaries defined by watershed) kind of an interesting thought experiment. At the very least it'd help to solve water politics.
They're absolutely gameable, and would lead to a lot of debate. Maybe we should work on standardizing and democratizing the redistricting process rather than just skipping to the end.
I'm just saying that I think we should take gerrymandering more seriously than pretending that it's a map coloring problem. There's certainly a lot of math involved, but the properties that are theoretically desirable in a district are not straightforward.
I fear this recipe (especially for the political party balancing) might be optimizing for strife. The only people who would win would be local TV/radio and direct mail in an endless election cycle.
I would argue that it’s more important for watersheds to align on jurisdiction boundaries. Representative government doesn’t need to fall on the same lines as local policy decisions on pollution rules, water rights, and emergency prep.
In the same way that some districts are richer than others, water is produced in some watersheds and consumed in others. Strongly aligning politics to those watersheds might not promote the cooperation you expect.
I do find watershed democracy (legislative district boundaries defined by watershed) kind of an interesting thought experiment. At the very least it'd help to solve water politics.