In theory, darkening a portion of the Earth with high albedo (snow, sand) is worse than darkening a portion of the Earth with low albedo (roads, roofs, forest). Then it should be better to use a greener area for solar panels so long as the capacity factors would be similar.
Also I don't see how that justifies adding to the effect? Especially if we don't know can't compute the trade off? I'm sure someone has done that work but my it didn't immediately jump out at me during a cursory search.
Inventing BS problems does not contribute to the discussion. Please stick to facts.
This depends on the efficiency of the PV modules. When it's greater than the albedo of the desert, it will cool the desert, not warm it. Remember, the electrical energy produced isn't heat. It's turned to heat at the consumer, but that's true of electrical energy from a nuclear plant too. Oh, and did you know that nuclear plants also add 2x as much heat energy in the exhaust of their cooling system?
Would covering a desert in solar panels cause more thermal solar absorption in that area than would otherwise happen? Yes.
But if we're optimizing for "offsetting the heating effect of human GHG emissions", then installing 4.5TW of solar (about 4x what has been installed worldwide to date) would have a much more positive effect.
The world currently has 1.1TW of solar installed, producing about 6% of all electricity. So our new installation would be on its own capable of supplying 25% of global electricity usage. The corresponding drop in GHG emissions from the shutdown of coal, gas, and oil power plants would far outweigh the fact that part of the desert has been turned black.