r.watershed: optionally use already existing maps of flow-accumulation or drainage direction

[TillF]

Is your feature request related to a problem? Please describe.
Many of the outputs of r.watershed depend on the internally produced maps of flow-accumulation and drainage direction. For massive grids, their generation is especially time consuming. It would be awesome if these maps could be re-sued in subsequent calls of r.watershed, e.g. when trying different subbasin delineation, etc.

Describe the solution you'd like
Introduce a flag "use_existing_rasters", which indicates that any maps accumulationor drainage that are found will be used instead of being recalculated.

Describe alternatives you've considered
Re-running r.watershed is possible, but computationally demanding.
Alternatives to r.watershed (e.g. r.terraflow) have the same limitation.

[petrasovaa]

Would any of the r.stream.* tools work for you? They can take external accumulation.

[TillF]

Thanks for this idea.
Specifically, I need to produce the outputs "basin" and "half_basin" of r.watershed, but with different thresholds.
My current workflow requires to have "basin" first, then produce "half_basin" later.
As far as I understand, with r.stream.basin this would need to be reversed, as it can only produce "basin" . Or did I oversee an option here?

[sumitchintanwar]

I am thinking that I create two additional optional parameters, when both these parameters are given, search will be skipped and instead of recalculating, it will load the existing accumulation and drainage values directly from the input maps and jump straight to the watershed basin determination step.

This will make re-runs nearly instantaneous for parameter tuning as shown in my local testing.

I will open a pull request shortly with this fix.

[sumitchintanwar]

This issue is addressed in PR #6953. Please let me know if any changes are needed. I’d appreciate your review and feedback.