Baur et al. evaluate the potential for solar photovoltaic and concentrating solar power under three future climate scenarios: SSP5-8.5, SSP2-4.5, and G6sulfur which reduces the climate forcing from SSP5-8.5 to SSP2-4.5 using stratospheric aerosol injection (SAI). They find the resource potential for both technologies reduces under the geoengineering scenario. The results confirm the one study that has previously investigated SAI impacts on solar energy technologies, from Smith et al. (2017).

The study is a development over Smith et al. (2017) in two regards. Firstly, the authors consider locational feasibility of solar power installations, ruling out or downweighting grid cells that are in protected areas, far from population centres, and conflict with existing land use types. The second is that the authors consider the intra-year variability in solar energy resource, referring to “low energy weeks” in which meteorological conditions do not produce sufficient energy. I also quite like that the authors used hourly data output from the climate model (compared to three-hourly from Smith et al.). With these additions, the results are similar to Smith et al. (2017), indicating robustness in the (admittedly intuitive) statement that SAI leads to reduced solar energy potential. Given the increasing occurrence of SAI in policy discourse, it is important that studies like these get a renewed focus, as the negative impacts on conventional mitigation (e.g. renewable technologies) of geoengineering are often not considered.

Main comments:

1. There does not appear to be a consideration of the solar geometry in the equation for PV. For CSP, the factor of the cosine of the zenith angle cancels out (Smith et al. eq. (7), eq. (9)) so providing the FLH equation is correctly defined in your paper then this is OK. However for PV, the direct/total irradiance is important, as well as the orientation of the solar panel, as to the amount of radiation it receives and the panel temperature which affects its efficiency. In eq. (1), the power output expected would be greater than predicted from the climate model value of RSDS, since this would be a horizontal irradiance value, and a real-world solar plant operator would angle the panels appropriately to maximise the incident irradiance on the panel. Perhaps these corrections are already baked into the equations you use. It would be good to confirm.
2. Around line 272 there is a ”quasi-linear” relationship for reduced potential. It looks fairly linear in time (fig. 6), but since we don’t have the SAOD plot we don’t know if it is linear in AOD. This would be quite a useful result to verify as if it is linear, it would be easy to transplant into an economic or integrated assessment model.

Minor points

Lines 35-37: several of the references are repeated

Line 73: “dystopian” I suppose is a slight value judgement

Line 86: A brief descripton of what G6sulfur aims to do, and the experiment design, would be useful.

Line 87: “imitates”: I take from this that CNRM-ESM is not emissions driven for stratospheric aerosol injection. It is mentioned in the discussion, but would be good to introduce here. Related to my comment about experiment design, what is the total loading or optical depth required to achieve the desired avoided warming?

Line 92: “aerosol-light interaction”, do you mean “aerosol-radiation interaction”?

Equations 1 and 4: the LHS looks like a subtraction, would be better to be a subscript TPi

Line 108 & 133: Format -2 superscript

Line 134, and a few other places – reference formatting a little sloppy and haphazard

Line 157: is there a basis for choosing 500 km as the cut-off?

Line 169: are low energy week boundaries fixed (i.e. Monday to Sunday), or do you take 7-day rolling averages?

Line 183: do the different population masks significantly affect the results? It feels like this isn’t quite an apples to apples comparison.

Line 208: delete first “in”

Line 275: “a lot less well-behaved” – being a bit more explicit/formal here is useful.

please see the attachment.