Shape-agnostic wind and solar potentials

[timtroendle]

What can be improved?

We currently use wind and solar potentials from Zenodo that are aggregated to shapes on the continental, national, and regional layer. This is a fast and pragmatic approach, but it's also inflexible:

First, it does not allow to easily build in other resolutions without updating the input data (see for example #319).
Second, it makes filtering the scope on the continental resolution complicated (see #252).

A more flexible, even if both computationally and storage-wise more intense, approach would be to use raster data as an input and then cookie-cut shapes from that data.

Version

1.2.0.dev

[timtroendle]

I never published the raster data, but I still have access to the original raster data from November 2019. There should be no difference in that data compared to the one we currently use. I can upload that data to Zenodo.