Hi gb

The path-length for light-rays through the atmosphere - if we throw in the stratosphere to 50 km - is over 800 km when the sun sits on the horizon. That's 16 times the zenith distance. Energy loss through an absorbing/scattering medium goes like so...

(I/Io)^(z/zo)

...where I, Io are the intensity and the initial intensity respectively, and z, zo are the path-length and initial path-length. Thus if 50 km attenuates the intensity to 75%, 16 times that reduces it to just 1%. At 30 degrees above the horizontal for a Sun path through the zenith and the intensity is 50% of its zenith value. Some PVs won't work below about 40% of their normal operating light levels and thus are only good for about 70% of day-time even when pointing straight at the sun. That's an operating voltage issue which can be overcome by concentrators for some configurations, but enough clouds can cause most PVs to be utterly useless and are no joy for solar thermal systems at all.

To compensate for 50% cloud cover during a year and maybe 40% useful sunlight on a good day, to supply continual power a ground based PV system needs big storage batteries of some sort - vanadium flow batteries are basically unlimited in size and easily transportable, thus are a fantastic option. To provide a continual energy flow the system would need to be anything from 5 to 10 times the baseload level at peak output, to account for storage inefficiencies and power supply discontinuity.

So of course SBSPs are already out in front because the supply is continual - except for hemi-annual eclipses for an hour or so.