Yes, I agree that could reflect in some way, but what seems strange is that the shape of the brightness doesnt seems in any case with the shape of the roof, and of anything around. The example you show of the solar panels is strictly the same shape of the real material, in summer, and winter, with and without reflection.
I used to work with Fotogrametry and GIS. I worked for several years with cartography, satellite images...
I asked some Doctors, and proffessionals and they said that is strange. It seems like a post image edition. If you look pretty close you can see.
hi cespombo ! my previous screenshots were not example for flares, only for that even black surfaces can appear bright on aerial or satellite imageries.
flares (on satellite imagery) nothing to do with original shape of the reflecting object, only it needs perfect perpendicular angle of the lightbeam which arrives back to satellite sensors more intense, than it could handle by processing and the result is a distorted, shapeless image error what appears on imagery.
think, I could dig up some proof for you:
similar looking glitch to yours from Vien: flare Wien.kmz (807 B) (note, flare occured at mid summer, right at noon as imaging satellite was at zenith so circumstances were ideal. - shape of flare is rather similar to yours, only it is upside down. - I am not sure whether the explanation for this could be that it is on the other hemisphere, or that the satellite moved the opposite direction. - also can be simple coincidence.)
closer view by different imagery date to show, what possibly caused the "flare" in this case: flare source Wien.kmz (810 B)