I agree with regards to rooftop being an important part, though. It makes sense to use land that is already occupied to a large degree.
Insolation: 1 kW/m^2
Panel efficiency: 10-20%, nominally 17%. Maximum for a single-layer cell is 37% for quantum physics reasons, with an "infinite" layer panel, 87%.
Capacity factor is the amount of time a panel actually recieves sunlightlight. Typically 20-30%, with 30% as a peak.
Spacing factor. PV doesn't work if shaded -- the entire circuit cuts out. So you have to angle and space your panels. On net, this reduces availble area by about 50%. Less near the equator, more at high latitudes.
Degredation. For numerous reasons, solar panels degrade with time, mostly due to physical wearing (broken cells, wiring, hazing of panels, stone and hail damage, etc.) When nominal output falls to about 85% new, panels are deemed due for replacement. This is about 20 years based on multiple studies. So you're replacing 5% of your stock every year, and losing a few percent of output.
Inverter losses. Converting panel (DC) to transmission (AC) power costs about 10%.
Transmission losses. Another 6% or so for high-voltage transmission.
Multiply all that out, and your 1 kW/m^2 * 0.17 * 0.50 * 0.9 * 0.9 * 0.25 leaves you with 17 watts of 24/7/365 capacity per square meter. Or about 1.7% net efficiency.
This doesn't account for storage which can cut into that some more.
Does that sound like a lot to you? A single airport runway is much larger than 1 sq km.
Point: a square kilometer to generate 1GW peak of electricity would be an obviously excellent trade off, well within the cost parameters that humans are willing to spend on far less useful endeavors like runways and parking lots.
I picked Midway because it's a very conveniently sized airport, being almost exactly 1 mi² in a square-shaped with obvious meter marks (due to surrounding grid system).
I've heard & use the term solar strip-mining.
Solar power maxes out at no better than 1kW/m^2 (sun at zenith, clear day, practically 100% efficient conversion), we're talking an absolute minimum of 1 sq km to match a baseline 1GW power plant ... and likely 100 times larger for real-world scenarios (10% efficiency, night/clouds, non-zenith air densities, dust, etc). So yeah, the net effect is "strip-mining" a vast area by removing all (or a seriously impactful percentage) natural power input and thus killing everything off.
Rooftop solar should indeed be the focus. There's even solar panels intended as road/pavement surfaces, super-durable interlinked hexes grabbing power where we've already wrecked the natural surface.