Allow higher capacity_per_sqkm for solar technology in deserts

[pz-max]

Allow higher capacity_per_sqkm for solar technology in deserts

We are currently defining how much solar can be installed in a region by a parameter in the config: https://github.com/pypsa-meets-africa/pypsa-africa/blob/d86fed5a6dee9e6e55b9ceed6ef20042ce6fab16/config.default.yaml#L220 In another issue https://github.com/pypsa-meets-africa/pypsa-africa/issues/361 we compare the assumption and investigate the impact to the technical potential.

The parameter was chosen to represent socio-technical and political limits. The current value in PyPSA-Eur is 1.7 MW/sqkm. The average power density of utility-scale solar power plants is around 45.6 MW/sqkm (confirmed by looking at 41 power plants in a notebook that we upload soon). Maximal observed densities are even around 100sqkm. PyPSA-Eur's current default says that roughly 3.7% (~1.7/0.037=46) of all available areas (which already exclude protected areas and specific landcover codes) can be used for capacity expansion.

Obviously, one can criticize if 3.7% of all available areas is really the socio-technical and political limit or not. We will actually be a bit optimistic and assume that 10%, representing 4.6 MW/sqkm, of all available areas can be used for solar expansion (new solar design also allows synergies between farm & cropland reducing such limitations, see agrivoltaics).

But is this socio-technical and political assumption applicable to deserts (Copernicus LC class 60)? I think no. I would argue that we should do the following:

• Using for different land-cover classes different capacity_per_sqkm assumptions. Here, I think the main impact in Africa would be to use the full average power density for deserts (45.6 MW/sqkm). There are simply no socio-techno and political limitations...

A pull request should update the solar capacity_per_sqkm for Copernicus class 60 (e.g. desert). The code snippet that might be relevant: https://github.com/pypsa-meets-africa/pypsa-africa/blob/d86fed5a6dee9e6e55b9ceed6ef20042ce6fab16/scripts/build_renewable_profiles.py#L346-L353

Please give a clear and concise description and provide context why the feature would be useful. Also, we'd appreciate any implementation ideas and references you already have.

[ekatef]

An updated link to Copernicus land coverage codes: https://land.copernicus.eu/en/technical-library/global-dynamic-land-cover-product-user-manual-v3.0/@@download/file

FYI @GbotemiB

[ekatef]

A recent work of PV/CSP capacity: Franke, Katja and Fragoso, Joshua and Kleinschmitt, Christoph and Sensfuß, Frank, Assessing Worldwide Future Potentials of Renewable Electricity Generation: Installable Capacity, Full Load Hours, and Costs. Available at SSRN: https://ssrn.com/abstract=4526901 or http://dx.doi.org/10.2139/ssrn.4526901