Secondly, what genius decided to make it a flat roof? Flat surfaces are the absolute worst both in capturing the most sunlight in temperate latitudes and in terms of dirt and snow removal. Even a 45° angle would mean a lot of the dust, leaves and snow will just tumble down or get washed down by the next rain, whereas flat surfaces are literally encrusted within months.
You want solar, put it in an uninhabited ground-level unobstructed serviceable cheap location near the grid outside of town. Because, wires can transmit electricity anywhere it's needed.
Elaborate structures for solar cost more than the panels these days. Never build them. At most, tilt them to the local lattitude. If tempted to do more, instead buy more panels. Power per dollar works out better.
They probably did the math, and at that location, the ROI for a single-axis tracker was better than the ROI for fixed panels.
> And they track over the day! not over the year.
This probably means it's a single-axis tracker, which points east in the sunrise and west in the sunset. To track over the year it would need a second (north-south) axis, but the ROI for that probably wasn't better than a single-axis tracker.
> Again, all that mechanism for so little gain.
A single-axis tracker isn't that complicated (and a single axis with a single motor can move a whole row of panels at once), so even if the gain is small, it can be more than enough to pay for the mechanism.
They've gotten so cheap that this Texas company claims that it's cheaper to buy more panels and lay them on the ground rather than using a rack to get them at a better angle.
https://electrek.co/2022/12/12/texas-solar-farm-flat-on-the-...
They probably didn't do the math. If they had they would have installed fixed panels. Panels are so dirt cheap that you'd have to come out more than 60% ahead on the cost of the tracker and that's never ever going to happen. Just add more panels.
I have 50 of them now in some pretty bad orientations, ~ 100 KWh on a good day and if I had half that many on trackers it wouldn't even come close. And on bad days the only thing that you want in a solar installation is surface area. Peak performance isn't interesting at all, and tracking at best gives you 20 to 25% more power at the price of the tracking mechanism being much more bulky than panels, panels shading each other when used with trackers so you'll use your space far less efficiently, fragility and finally complexity. Don't use trackers.
This method adds no land use. It also provides a roof for the users of the path.
Solar above car parks is great too -- it provides shade, means you don't get wet getting out of the car, and it provides energy.
You want a roof, build a roof. You want solar, build a solar farm. You should no more put solar over other structures than grow radishes up there. It's a terrible idea from every point of view (except eco-theatre).
> wires can transmit electricity anywhere it's needed.
Well .. yes, if you build the wires. A surprising amount of projects are getting held up on grid connection grounds. Whereas embedding the panels near urban demand reduces that problem.